Design and Development of a Myoelectric Transradial Prosthesis

Design and Development of a
Myoelectric Transradial Prosthesis
A Major Qualifying Project Design Report:
Submitted to the Faculty of
WORCESTER POLYTECHNIC INSTITUTE
In partial fulfillment of the requirements for the
Degree of Bachelor of Science
By
________________________
Cameron Currie (BME/ME)
________________________
Cameron Downey (ME)
________________________
A. Rae Nistler (ME)
________________________
David McDonald (ME)
________________________
Gregory Port (RBE)
________________________
Joseph Sabatino (ME)
________________________
Steven Souto (ECE)
Date:
Approved:
_____________________________
Prof. David C. Planchard
Advisor (ME)
_____________________________
Prof. Tiffiny Butler
Co-Advisor (BME)
_____________________________
Prof. Sergey N. Makarov
Co-Advisor (ECE)
_____________________________
Prof. Craig Putnam
Co-Advisor (RBE)
This report represents the work of Worcester Polytechnic Institute (WPI) undergraduate
students submitted to the faculty as evidence of completion of a degree requirement. WPI
routinely publishes these reports on its website without editorial or peer review.
For more information about the projects program at WPI, please see
http://www.wpi.edu/academics/ugradstudies/project-learning.html.
Abstract
The loss of a limb is a life-changing event and reality for 441,000 transradial amputees in
the United States. Limb loss can have substantial physical, social, psychological, and economic
consequences. A prototype prosthesis was created that has sophisticated hand functionality, an
adjustable and comfortable socket, and a lightweight yet durable design utilizing 3D printing, all
available at a reasonable price point. The prosthesis integrated force sensors, servo motors, and
a myoelectric means of control so the user may perform activities of daily living. The overall
outcome was a prosthesis that met its design requirements, offering increased usability,
functionality, and availability.
1
Table of Contents
Abstract
1
1. Introduction
7
2. Background
8
2.1 Amputations and the Goals of Prosthetics
8
2.1.1 Amputations and Their Effects
8
2.1.2 Goal of Prostheses
9
2.2 Anatomy and Physiology of the Forearm
10
2.2.1 Anatomical Position
10
2.2.2 Bones
11
2.2.3 Joints
12
2.2.4 Muscles
13
2.2.5 Anatomical Change from Transradial Amputations
13
2.2.6 Lost Articulation
15
2.3 Transradial Prostheses
16
2.3.1 Current Issues
16
2.3.2 Types
16
2.3.3 Components
17
2.3.4 Functionality
17
2.3.4.1 Hand Grips
18
2.3.4.2 Fingers and Their Articulation
20
2.3.5 Materials and Manufacturing Processes
2.4 Current Technology
22
23
2.4.1 Open Bionics’ Ada Hand
23
2.4.2 e-NABLE’s Limbitless Arm
23
2.4.3 InMoov
24
2.4.4 Rehabilitative Robotic Glove
24
2.4.5 Multi-DOF Anthropomorphic Prosthetic Hand
25
2.4.6 Design of a Human Hand Prosthesis
26
2.4.7 Socket
27
2.4.7.1 Current Socket Technology
27
2.4.7.2 Vacuum Suspension Sockets
28
2.4.7.3 Bladder-Controlled Socket
29
2
2.4.7.4 Compression/Release Stabilized Socket
30
2.4.7.5 Exact Limb Replica
31
2.4.7.6 Mechanically Adjusted Socket
31
2.5 Impacts and Ramifications
32
2.5.1 Manufacturing, Sustainability and Environmental Aspects
32
2.5.2 Ethical Concern and Health and Safety Issues
33
2.5.3 Economics, Political, and Societal Impacts
33
3. Methodology
34
3.1 Design Goals
34
3.2 Design Specifications
37
4. Design
4.1 Hand
39
39
4.1.1 Finger Design
39
4.1.2 Palm Design
41
4.2 Forearm
41
4.3 Socket
43
4.4 Motor Selection
47
4.5 Control System
47
4.5.1 DC to DC Converter
48
4.5.2 Myoelectric Input Block
48
4.5.3 Grip Selection
49
4.5.4 Actuation Feedback Block
51
4.5.5 Servo Control Block
51
4.5.6 Microcontroller
52
4.5.7 Battery Selection
54
4.5.8 Myoelectric Data Processing
55
5. Testing and Results
5.1 Tests
57
57
5.1.1 Pin Fracture
57
5.1.2 Thread Ductility
57
5.1.3 Hand Friction
57
5.1.4 Socket Slippage
58
5.1.5 Finger Actuation
58
3
5.1.6 Control System Hardware
58
5.1.7 Control System Software
58
5.1.8 Hand Functionality
59
5.2 Test Results
59
5.2.1 Pin Fracture
59
5.2.2 Thread Ductility
59
5.2.3 Hand Friction
59
5.2.4 Socket Slippage
59
5.2.5 Finger Actuation
60
5.2.6 Control System Hardware
60
5.2.7 Control System Software
60
5.2.8 Hand Functionality
60
6. Project Expansion and Continuation
61
6.1 Haptic User Feedback
61
6.2 Cooling
61
6.3 Wrist
63
6.4 Scalability
64
6.5 Engineering Standards
65
References
66
Appendices
72
Appendix A: Nomenclature
72
Appendix B: Complete Directional Terms in Human Anatomy
73
Appendix C: Activities of Daily Living
75
Appendix C.1: Detailed Descriptions of the Activities of Daily Living
75
Appendix C.2: Hand Usage during Activities of Daily Living
75
Appendix D: Design Goals Decision Matrix
76
Appendix E: Schematic for Control System
77
Appendix F: Full Test Procedures
78
Appendix F.1: Pins Test Procedure
78
Appendix F.2: Finger Actuation Test Procedure
83
Appendix F.3: Thread Ductility Test Procedure
87
Appendix F.4: Socket Slippage Test Procedure
90
Appendix F.5 Control System Hardware Test Procedure
94
Appendix F.6 Hand Functionality Test Procedure
103
4
Appendix G: Model Prototype Images
105
Appendix G.1 Bottom View
105
Appendix G.2 Isometric View
106
Appendix G.3 Manufactured Design
107
List of Figures
Figure 1: Amputation Statistics ................................................................................................... 9
Figure 2: Activities of Daily Living ............................................................................................... 9
Figure 3: Anatomical position and planes of reference ..............................................................10
Figure 4: Bones and joints of the hand ......................................................................................12
Figure 5: Muscles of the hand ...................................................................................................13
Figure 6: Muscles of the right forearm .......................................................................................13
Figure 7: Transradial amputation ..............................................................................................14
Figure 8: Surgical process of transradial amputation .................................................................14
Figure 9: Wrist, Finger, and Thumb Articulation ........................................................................15
Figure 10: MyoWare Muscle Sensor .........................................................................................16
Figure 11: Power Grips .............................................................................................................18
Figure 12: Precision Grips .........................................................................................................19
Figure 13: Underactuated Linkage around Spherical and Irregular Objects...............................20
Figure 14: Underactuated linkage design (Multi-DOF Anthropomorphic Prosthetic Hand) .........20
Figure 15: Mechanical Linkage Design .....................................................................................21
Figure 16: Flexible Finger Joints ...............................................................................................21
Figure 17: The Ada Hand ..........................................................................................................23
Figure 18: Limbitless Arm .........................................................................................................23
Figure 19: InMoov Hand ............................................................................................................24
Figure 20: Rehabilitative Robotic Glove ....................................................................................24
Figure 21: Three motor design for anthropomorphic hand .........................................................25
Figure 22: Control System for Force Feedback Controlled Finger .............................................25
Figure 23: Mechanical Linkage for Finger Actuation ..................................................................26
Figure 24: Finger Actuation System ..........................................................................................26
Figure 25: Photos of residual limbs suffering from skin problems ..............................................27
Figure 26: Vacuum suspension socket and Inflatable bladders .................................................29
Figure 27: Compression/Release Stabilized Sockets ................................................................30
Figure 28: 3D Printed Socket Creation Process ........................................................................31
Figure 29: The Revolimb and the Socket-less Socket ...............................................................32
Figure 30: SolidWorks Model of Hand Design ...........................................................................39
Figure 31: Push Pull Pulley Design ...........................................................................................40
Figure 32: Fully Actuated Finger ...............................................................................................40
Figure 33: Palm Design.............................................................................................................41
Figure 34: Palm Cover ..............................................................................................................41
5
Figure 35: Render of Full Assembly ..........................................................................................42
Figure 36: Servo Frame, without Servos ...................................................................................42
Figure 37: Socket Frame CAD Model ........................................................................................44
Figure 38: Base Plate-Hinge Connectors’ Positions for Girth Adjustments ................................44
Figure 39: Base Plate-Hinge Connector Angle Adjustment .......................................................45
Figure 40: Hinge Angle Adjustment ...........................................................................................45
Figure 41: Small, Medium, and Large Strut Lengths .................................................................46
Figure 42: Non-Amputee Showing Strut Positioning ..................................................................46
Figure 43: Hitec HS-5585MH ....................................................................................................47
Figure 44: Converter Schematic ................................................................................................48
Figure 45: 6-pin configuration....................................................................................................49
Figure 46: Grip Selection Block Schematic ...............................................................................50
Figure 47: Force Sensor Array Schematic.................................................................................50
Figure 48: Sensor Placement ....................................................................................................51
Figure 49: Servo Control Block .................................................................................................52
Figure 50: Microcontroller .........................................................................................................53
Figure 51: Logic Flow Chart ......................................................................................................53
Figure 52: Kalman Filter on Muscle Sensor Data ......................................................................55
Figure 53: Results of Early EM Algorithm Prototype on Filtered Sensor Data ...........................56
Figure 54: Possible design for wrist...........................................................................................64
Figure 55: Equations and relations table ...................................................................................65
List of Tables
Table 1: Directional Terms in Human Anatomy
Table 2: Average Grip Strengths
Table 3: Design Specifications
Table 4: Socket Frame Decision Matrix
Table 5: Finger Positions
Table 6: Estimated Battery Life Based on Servo Power Draw
Table 7: Cooling System Decision Matrix
11
19
37
43
49
54
62
6
1. Introduction
The loss of a limb is a life-changing event and a reality for over 2.1 million people in the
United States. Of those 2.1 million, 441,000 are transradial, below the elbow, amputations. Upper
limb loss has substantial physical, social, psychological, and economical consequences for an
amputee. In order to mitigate these consequences and assist the amputee to return to a state of
normalcy. Transradial prosthetics, or artificial hands and wrists, are used to perform daily activities
such as eating and dressing. The purpose of transradial prosthetics is to help the amputee
function as nearly as well as before.1
The purpose of these prosthetics hasn’t changed; however, innovation in technology has
vastly improved their performance. Unfortunately, advanced technology has its price. State-ofthe-art prosthetics are extremely expensive at hundreds of thousands of dollars. Even cosmeses,
prosthetics made only for aesthetics, although less expensive, are still thousands of dollars.
Innovations in 3D printing, an additive manufacturing process, over the past decade have
made 3D printed transradial prosthetics an inexpensive alternative. They are able to provide
complex functions at an affordable price. Multiple non-profit organizations have independently
formed to create innovative 3D printed transradial prosthetics. There are multiple designs that
range in ability and function such as scalability, durability, hand grips, control systems, user
inputs, materials, aesthetics, comfort, and cost. Each prosthetic design has its own strengths and
weaknesses.
This project identified areas for improvements and strengths of these 3D printed
transradial prosthetics. The team combined the strengths of many prostheses with their own
innovative ideas to design a 3D printed transradial prosthetic. A prototype prosthesis was created
that has sophisticated hand functionality, an adjustable and comfortable socket, and a durable
yet lightweight design utilizing 3D Printing, all available at a reasonable price point. The prototype
was then evaluated against the required specifications. The subject of this report is the design
process, manufacture, and testing of the 3D printed transradial prosthetic.
1
Advanced Amputee Solutions, 2016
7
2. Background
In order to create an appropriate design, amputations’ frequency, causes, demographics,
types, and their negative effects as well as the impact prosthetics have on alleviating these
negative effects were investigated. Next, the anatomy and physiology of the forearm and an
amputated forearm were explored in order to more easily convey later design intentions. The
purpose, components, functions, and general design of modern transradial prostheses are
described, and several examples of existing technology are analyzed in the following.
The team’s goal involved building upon preexisting research in many areas of prosthetics.
Several different prosthetic arm/hand designs were analyzed and their best and worst qualities
were recorded. In addition, the more advanced systems of prosthetics such as finger actuation,
wrist articulation, the connection between the residual limb and the prosthetic, also known as a
socket, and the control system of the entire prosthetic were analyzed as well. The team looked at
various methods for implementing each of these systems before making design decisions on their
inclusion in the final prototype.
2.1 Amputations and the Goals of Prosthetics
The overall goal of prosthetics is to help normalize amputees. Prosthetics accomplish this
goal by returning functions that their lost limb previously provided. For example, a prosthetic leg
restores the ability to walk; a prosthetic heart valve replaces a damaged, natural valve and allows
for better blood flow in the heart; and a prosthetic arm gives an amputee the ability to once again
manipulate their surroundings in a more “normal” fashion.
2.1.1 Amputations and Their Effects
Amputation is the last resort when surgical salvage is not possible. Upper extremity
amputations can have substantial physical, psychological, social, and economic consequences
for the patient.2 Thirty-six percent of amputees living with limb loss suffer from depression.3
Amputees pay over half a million dollars in healthcare costs over their lifetime. In the United
States, hospital charges for patients who undergo amputations totaled to $8.7 billion in 2013.4
Surgeons try to mitigate consequences by providing maximum use of the residual limb
without a prosthesis and minimizing the known complications of amputation. The ultimate goal of
amputation surgery is to provide a sensate limb that can best interact with the patient’s
environment, with and without a prosthesis.5
There are over 2.1 million amputees currently living in the United States. That number is
expected to grow to 3.6 million by 2050.6 Each year, 185,000 people have an amputation. This
equates to 507 people losing a limb each day. The two most common causes of these
Marchessault, McKay, & Hammert, 2011
Advanced Amputee Solutions, 2016
4 Ibid
5 Marchessault, McKay, & Hammert, 2011
6 Advanced Amputee Solutions, 2016
2
3
8
amputations are vascular disease (54%)
and trauma (45%).7 46% of amputations,
happen between the ages of 45 to 64,
however limb loss affects people of all
ages, from birth to over 85 years old.
Men are approximately two times
more likely to have an amputation than
women. Upper limb amputations account
for 35% of all amputations.8 Of these upper
limb amputations, transradial amputations
comprise 60%.9 This means, nationally,
approximately 441,000 people are living
with a transradial amputation.
2.1.2 Goal of Prostheses
Figure 1: Amputation Statistics
Prostheses rehabilitate amputees
by restoring as much function as possible.
The devices do this by targeting functions that fulfill Activities of Daily Living, or ADLs. ADLs are
routine activities that people tend to do every day without needing assistance. There are
approximately eight10 ADLs:
1.
2.
3.
4.
5.
6.
7.
8.
Food Preparation
Feeding
Personal Care
Housekeeping
Shopping
Driving and Transport
Leisure
Others
For detailed descriptions of these
ADLs, including a breakdown of how much
time is spent on each task on average, see
Appendix C. In upper limb prostheses, the
restoration of ADLs is essential, as almost all
activities demand the use of one or more
hand. Approximately five hours out of the day
alone is spent by hands completing these
essential activities.11 This amount of time
Figure 2: Activities of Daily Living
Amputee Coalition, 2016
Ibid
9 Ziegler-Graham et. al, 2008
10 Vergara, M. et al., 226
11 Ibid
7
8
9
does not take into consideration all of the non-essential activities the human hand and arm
participate in each day.
2.2 Anatomy and Physiology of the Forearm
It is necessary to understand the basic anatomy and physiology of the human hand in
order to successfully design a prosthetic device. Anatomy is the study of structure, and physiology
is the study of function. Through the study and understanding of the structure and function of the
human hand, it can be mimicked for a better prosthesis.
2.2.1 Anatomical Position
When discussing anatomy and physiology, there are certain assumptions and terminology
about the position of the body that are made. The anatomical position of a person is considered
to be standing erect with feet flat on the floor and close together, arms at the sides, and the palms
and face directed forward as shown in Figure 3. This provides a frame of reference in order to
discuss the details of the human body. Table 1 includes the most relevant and commonly used
position related terminology and their definitions. For a complete list, see Appendix B.
Figure 3: Anatomical position and planes of reference12
12
Saladin, 2012
10
Table 1: Directional Terms in Human Anatomy
Term
Definition
Sagittal Plane
Passes vertically through the body. Divides the body into right and left
sections
Median Plane
Sagittal plane that divides the body into two equal halves
Frontal Plane
Passes vertically through the body, but is perpendicular to the sagittal plane.
Divides the body into anterior and posterior sections
Transverse
Plane
Passes horizontally through the body. Divides the body into superior and
inferior sections
Anterior
Toward the ventral side
Posterior
Toward the dorsal side
Superior
Above
Inferior
Below
Medial
Toward the sagittal or median plane
Lateral
Away from the sagittal or median plane
Proximal
Closer to the point of attachment or origin
Distal
Farther from the point of attachment or origin
Superficial
Closer to the body surface
Deep
Farther from the body surface
2.2.2 Bones
Transradial prostheses replace bones in the hand, wrist, and forearm. Bones provide
structure and support that transradial prostheses mimic in the structure and casing. The human
11
hand is comprised of five different sets of bones, 27 bones overall. The five sets are the carpals,
metacarpals, first phalangeal, second phalangeal, and third phalangeal bones. Each finger
contains three phalanges, with the exception of the thumb which only has two. The phalanges
bones are known as the distal, middle, and proximal phalanges. The thumb only contains the
distal and proximal phalange.13 There are five metacarpal bones in the palm, which are numbered
one through five starting from the thumb moving to the pinky finger. The eight carpal bones make
up the wrist and are arranged in two rows. The ulna and radius are the bones of the forearm. The
radius bears about 80% of the force on your forearm while the ulna shares the load and minimizes
wear and tear. The bones of the hand can be seen in Figure 4.
Figure 4: Bones and joints of the hand14
2.2.3 Joints
Transradial prostheses replace joints in the hand and wrist. Joints provide flexibility, range
of motion, and the ability for articulation that transradial prostheses mimic using mechanical joints.
Transradial prostheses typically use pins, hinges, and ball joints. The hand contains six different
joints, the distal interphalangeal (DIP), proximal interphalangeal (PIP), interphalangeal (IP),
metacarpophalangeal (MCP), carpometacarpal (CMC), and radiocarpal (RC) joints. Each digit of
the hand contains the DIP, PIP, and MPC which all have ligaments that provide stability. The
thumb, however, contains the IP instead of the DIP and PIP as shown in Figure 5. The
trapeziometacarpal (TMC) joint at the base of the thumb contains the clearest example of a saddle
joint. Saddle joints are biaxial and have greater range of motion than the other phalanx’s condylar
joints. This range of motion gives humans an opposable thumb. The CMC is between the
metacarpals and the carpals. This joint allows the hand to curl and grasp objects. The radiocarpal
is the joint between the carpals and the radius and ulna. It enables the wrist to flex and extend.15
Saladin, 2012
Brigham and Women’s, 2014
15 Saladin, 2012
13
14
12
2.2.4 Muscles
Transradial prostheses replace muscles in the hand,
wrist, and forearm. Muscles provide movement that transradial
prostheses mimic using servos, control systems, gears,
hinges, and other components. The muscles that control the
fingers and thumb lie in the forearm and hand. There are two
groups of muscles that comprise the hand and wrist: intrinsic
and extrinsic muscles. The intrinsic muscles of the hand
provide precise finger movement and allow for each finger to
move independently. Intrinsic muscles are separated into four
groups: the thenar muscles, that act on the thumb; the
hypothenar muscles, that act on the little finger; the lumbrical
muscles that help the extension of the IP joints and the flexion
of the MCP joints; and the interossei, that allow for abduction
Figure 5: Muscles of the hand17
and adduction of the fingers as shown in Figure 5.16
The extrinsic muscles of the forearm are larger, longer muscles that run from the forearm
to the hand and provide strength. Two important extrinsic muscles are the flexor digitorum
profundus (FDP) and the flexor digitorum superficialis (FDS). These muscles are utilized when
repetitive work and additional strength are necessary.18 Figure 5 and Figure 6 show the location
of some of these muscles.
2.2.5 Anatomical Change from Transradial Amputations
Transradial amputations change the anatomical structure of the bones, muscles, tendons,
and nerves. It is important to understand
these changes when designing a
transradial
prosthetic
device.
The
structure of the residual limb affects the
size, length, control, and overall structure
of the prosthesis. Each residual limb is
unique. This is the biggest challenge of
designing
and
mass
producing
prostheses. The surgical process and
reconstruction of the limb are the key to
understanding the differences between
residual limbs.
When the decision is made to
amputate an upper limb, preservation of
Figure 6: Muscles of the right forearm19
the length and joint function are of
Saladin, 2012
Kelso, 2015
18 Ibid
19 Ibid
16
17
13
paramount concern. The upper extremity’s interaction with the surroundings depends on the
major joints to move the hand through space. An essential part of preservation is ensuring there
is not further bone shortening by making sure there is adequate residual soft tissue. The residual
soft tissue, particularly muscle, must provide adequate soft tissue coverage of the radius and ulna
to allow stump closure.20 Myodesis, anchoring of muscle or tendon to bone, of the deeper forearm
muscles to the radius and ulna provides stable bone coverage and prevents bone-on-bone motion
that can lead to complications such as fluid collection in the stump.21 Myoplasty, muscle to muscle
attachment, of the superficial flexor muscles to the extensor muscles must be placed on tension
to allow contraction of the muscles after closure. Soft tissue coverage of the radius and ulna with
tensioning of the muscles is accomplished with both myodesis and myoplasty.22
Figure 7: Transradial amputation23
Myofascial closure, enclosure of the muscle and its sheath of connective tissue, often
indicated for dysvascular tissue amputations, is not strong enough for muscle contraction and
should be performed only to help contour the remaining muscle bellies to enhance closure.
Contractions of the superficial muscle groups are essential to trigger myoelectric prosthetics.
Amputation six to eight centimeters proximal to the wrist joint allows for ample muscle coverage;
however, ten centimeters proximal is advocated for increased prosthetic options. Forearm
amputation at five centimeters distal to the elbow joint is the minimum amount of residual limb
necessary for a transradial prosthesis to fit. Although pronosupination, turning of the wrist, is lost
with more proximal transradial amputation, preservation of elbow motion is worthwhile. Transfer
Figure 8: Surgical process of transradial amputation24
Marchessault, McKay, & Hammert, 2011
Singh, Hunter, & Philip, 2007
22 Marchessault, McKay, & Hammert, 2011
23 Ibid
24 Versalius, n.d.
20
21
14
of the biceps tendon to the ulna should be considered to lessen the risk of flexion contracture with
proximal transradial amputations.25
Figure 8 shows a rendition of a transradial amputation. Figure 8-A begins with the incision.
Moving from anterior to posterior, Figure 8-B shows the development of superficial and deep
flexor mass as well as extensor muscles. Figure 8-C shows myodesis performed with sutures
passed through tendinous portions of muscle, through bone tunnels, and passed back through
muscle. Figure 8-D shows myodesis of superficial flexors and extensors sutured to each other
with some tension. Figure 8-E shows the muscles contoured with myofascial sutures to
accommodate tension-free closure.26
2.2.6 Lost Articulation
Transradial prostheses mimic basic functions or types of motion of the human
hand. We identified the most important of these for fulfilling ADLs.
● Wrist pronation and supination
● Finger flexion and extension
● Thumb flexion and extension
Figure 9: Wrist, Finger, and Thumb Articulation
25
26
Marchessault, McKay, & Hammert, 2011
Ibid
15
2.3 Transradial Prostheses
There are many transradial prostheses available all of which differ in design in terms of
their control system types, components, mechanical functions, materials, and manufacturing
processes. This section discusses the advantages and disadvantages of each.
2.3.1 Current Issues
There are currently many different prosthetics offered to patients. All have different
functions or aspects to them making them unique or better than other products. However, the
amount of people who have suffered from traumatic experiences are not looking for extremely
complicated prosthetics. Patients desire a chance to return to normalcy and have some
assistance in doing so. When they seek out a prosthetic and see one that can handle more weight
than an average person would have to lift there would be no need or want in getting that prosthetic.
Products are not designing around the needs of the patients and are only adding features they
feel would attract the buyers. The prosthetic design took into consideration who would be wearing
the device and what they would most likely be doing in the day. The average person would only
need to lift around 10lb of weight and need a handful of gestures to complete all the activities of
daily living. The designed prosthesis is simple to keep costs low however, was designed while
considering how everyday people would use it.
2.3.2 Types
There are two transradial prosthetic designs that are based upon the type of control
system. A control system “manages, commands, directs or regulates the behavior of other
devices or systems”. Prostheses utilize both mechanical and electrical controls. Some simple
transradial prostheses, such as hooks, use purely mechanical systems. The more technologically
advanced systems use a combination of both but rely mainly on electrical.
Mechanical inputs refer to a physical movement made by the user that directly powers the
prosthetic. Transradial prostheses use the mechanical function of a joint, typically the wrist or
elbow. The articulation of the joint in conjunction with a mechanical system, such as pulleys and
cables, opens and closes the fingers. Purely mechanical systems are simplistic and provide high
sensory feedback; however, they require gross limb movement for operation. Users must have
the required limb strength and range of motion necessary to
effectively operate the prosthetic. For the purposes of this
project, mechanical inputs will only be used to supplement
electrical inputs.
Most electrical inputs still require a physical movement
made by the user. However, after the initial movement the input
is measured electrically and becomes an electrical signal.
Typical electrical input methods include pushing a button with a
preset function or reading changes in myoelectric impulses
created by muscle tissues. Myoelectric refers to the electrical
Figure 10: MyoWare Muscle Sensor
properties of muscles. When a person thinks about flexing a
16
muscle, the brain sends an electrical signal to the muscle. The muscle starts recruiting motor
units, or bundles of muscle fibers that generate the force behind the muscles. The harder a muscle
is flexed, the more motor units are recruited to generate greater muscle force. The greater the
number of motor units, the more electrical activity the muscle produces. Sensors can be used to
measure this electrical activity and its changes during flexion and extension. The voltage
differential between flexion and extension can be measured and translated into digital signals.
These signals can be used as an input. For example, when the bicep is flexed, the voltage will
increase and the signal can trigger the prosthetic hand to close. Then when the bicep is extended,
the voltage will decrease and the signal can trigger the prosthetic hand to open.
The process of collecting myoelectric input ranges from very complex, sophisticated
systems that can accurately determine the user's exact intention for muscle movement. However,
much simpler options are available for systems that only require the detection of any sort of
muscle flex. These systems, such as the MyoWare Muscle Sensor, use a signal amplifier and
electrodes to output a raw signal which can then be read by a microcontroller.
2.3.3 Components
Transradial prosthetic devices use multiple components working together in order to
function. The components mimic their corresponding human body parts as closely as possible in
structure. The major components are the hand, forearm, and socket.
The hand has three distinct parts: the palm, fingers, and thumb. The palm is important for
structural integrity, housing smaller components, and supporting the fingers. It is often used to
house electronics and other components that are used to articulate the fingers. The structure
must be sturdy and durable so that it can support the fingers and loads of gripping different
objects. Prosthetic fingers are often made with two or three joints. The use of two joints strays
from mimicking the human hand. In comparison, the DIP at the tip of the finger is eliminated and
held stationary. Using two joints, allows for the fingers to support heavier load while three joints,
allow for greater dexterity. Mimicking the human thumb is one of the biggest challenges as it has
a saddle joint and nine muscles working together to move it. Imitation of this movement is limited
by the amount of space available to the designer and the complexity, expense, and quality of the
components and electronics necessary. Most prosthetic thumbs only move with two degrees of
freedom along two perpendicular planes.
The socket attaches the rest of the prosthesis to the residual limb, providing support and
stability. Its design must balance form and function. The form must fit and adjust to the residual
limb with minimal slippage. If it doesn’t, the amputee can experience pain, sores, blisters, and
severe health concerns can arise. The prosthesis will also feel heavy and cumbersome,
compromising mobility.
2.3.4 Functionality
Transradial prosthetic devices have multiple functions that work together to mimic their
corresponding human body parts as closely as possible in function. The most important function
a prosthesis seeks to restore is traditional hand functionality, specifically the ability to grip various
objects. Ideally, the prosthesis should also function such that the hand can be articulated and
17
rotated without excessive use of the shoulder, as
a human forearm can rotate independently of the
upper arm. A successful transradial prosthesis is
relatively aesthetically pleasing, and allows the
user to operate normally, without excessive
compensation to accommodate the device.
2.3.4.1 Hand Grips
A major and critical function of a
prosthetic hand is to grip assorted objects. A
Figure 11: Power Grips28
person uses many different types of hand grips
depending upon the size, shape, and weight of
the object, as well as the task that is to be completed. Types of grasps fall into two categories:
power and precision grips.27 Designing for the ability to achieve these grips is near-essential to
the development of a prosthetic hand device.
Power grips utilize a significant amount of force. The fingers flex around an object in one
direction, while the thumb flexes around in the opposite direction. This provides a counterforce to
keep the object in contact the palm and/or fingers. Power grips include cylindrical, spherical, and
hook grips, as well as lateral prehension if the thumb is adducted away from the fingers.
When a cylindrical grip is used all fingers are flexed around the object, which is usually at
a right angle to the forearm. The thumb is wrapped around the object, often overlapping the
fingers. When using a spherical grip, all of the fingers and the thumb are adducted around an
object, and unlike the cylindrical grip, the fingers are more spread apart. The palm of the hand is
often not involved. The hook grip involves the four fingers flexed around an object in a hook-like
manner. The MCP joints are extended, and the PIP and DIP joints are in some degree of flexion.
The thumb is usually not involved. Figure 11 offers visuals of these grips.
Precision grips require more delicate movement and positioning of the fingers. They tend
to hold the object between the tips of the fingers and the thumb. Precision grips involve the
intrinsic and extrinsic muscles, the thumb is abducted, and the palm and proximal joints don’t
move. They are used for fine movement and accuracy; for example, when objects are small or
fragile. There are four types of precision grips: pad to pad, also called pinch or palmar; tip to tip,
also called pincer; lateral prehension; and lumbrical.
27
28
Steinfeld, 1986
Behrens, n.d.
18
When using a pinch grip, the MCP and PIP of the fingers are flexed, the thumb is abducted,
and the distal joints of both are extended to bring the pad of the fingers and thumb together.
Lateral prehension is the pad of the extended thumb pressing an object against the racial side of
the index finger. This type of grip is often used to hold keys, paper, or thin objects. The lumbrical
grip, sometimes referred to as the plate grip, flexes the MCP and PIP joints, extends the DIP joint,
and the thumb opposes the fingers, holding the object horizontal. Again, visual representations of
precision grips can be found in Figure 12 below.
Table 2, adapted from Mathiowetz, V., et al. (1985), shows the average strengths of grasps
for males and females in four categories: Grip Strength, Tip Pinch, Key Pinch, and Palmar Pinch.
These values are averaged from participants aged 20 to 75+ and across both hands.
Table 2: Average Grip Strengths
Average Strength (lb.)
Male
Female
Grip Strength
98
58
Tip Pinch
16.5
11
Key Pinch
24
16
Palmar Pinch
23
16
Prosthetics try to closely mimic these grips. However, the ligaments that incorporate the
joints of the wrist and thumb are numerous. It is extremely challenging to reproduce the same
degrees of freedom (DOF). Prosthetics often have fewer degrees of freedom and restricted
Figure 12: Precision Grips29
29
Ibid
19
movement. Therefore, most prosthetics cannot accomplish all of the grips the human hand is
capable of achieving.
2.3.4.2 Fingers and Their Articulation
Figure 13: Underactuated linkage design (Multi-DOF Anthropomorphic Prosthetic Hand)
The method by which the fingers of a prosthetic are articulated is an important system as
it can have effects on the cost, longevity, and durability of any given prosthetic. Prosthetics
typically utilize one of two methods to articulate the fingers: a mechanical linkage system or a
pulley system.
Both systems can be implemented to be underactuated, meaning the total degrees of
freedom of the system can be greater than the amount of control inputs. The figure below shows
how the same linkage can adapt to different surfaces. The effect here was described well in a
report from the Harbin Institute of Technology’s Robotics Research Institute: “Before contact, the
finger behaves as a single rigid body in rotation round the pivot in base joint. When the proximal
phalanx [the segment of the finger closest to the palm] makes contact with the object, the proximal
phalanx stops, and the other two phalanxes begin rotating and closing on the object because of
the effect of the underactuated linkages mechanism.”
Figure 14: Underactuated Linkage around Spherical and Irregular Objects30
The challenge that arises from using a mechanical linkage design within a finger is the
space limitation, and from a budget standpoint, the ability to 3D print the linkage system and
30
Pierluigi, 2011
20
Figure 15: Mechanical Linkage Design
maintain functionality. A linkage design from the Harbin Institute manages the space limitation
very well with their finger design. The finger is driven by a DC motor with a planetary gearbox for
added torque and is based on four-bar linkage mechanisms meaning it is underactuated.
A finger design posted to Thingiverse had a simple mechanical linkage system with no
underactuated linkage meaning the finger overall has only one degree of freedom. The design is
an example of a mechanical linkage that is relatively functional and 3D printable.
Another popular method that is most common in budget prosthetics for finger actuation is
a pulley system. The concept uses cables internally threaded along separate tracks in the length
of each finger. The cable or thread can then be pulled to actuate the fingers by means of
mechanical or motor input. The pulley system was used in the 2016 Worcester Polytechnic
Institute (WPI) Major Qualifying Project (MQP) titled Design and Manufacture of a Scalable
Prosthetic Hand through the Utilization of Additive manufacturing as seen in Figure 16 below. In
addition to this MQP, other previous designs include more than one cable, inflexible cable, or
electrical actuation.
Figure 16: Flexible Finger Joints31
31
ieeexplore.ieee.org
21
2.3.5 Materials and Manufacturing Processes
Prosthetics are made from a variety of materials. Some materials that are commonly used
for prosthetics are: carbon fiber, carbon plastic with a low melting temperature so the plastic can
be heated and free-formed, or bent, to better fit the user, steel, aluminum, polylactic acid (PLA),
Acrylonitrile Butadiene Styrene plastic (ABS), and other plastics such as polyethylene.
The material is very dependent upon the manufacturing process as all materials are not
able to be manufactured the same way. One type of manufacturing is rapid prototyping, a subset
of additive manufacturing. Rapid prototyping is a group of techniques used to quickly fabricate a
model of a part or assembly using 3D computer aided design (CAD) data. This concept fabricates
prototypes in a fraction of the time and cost when compared to traditional methods. There are
many rapid prototyping technologies. One of these technologies is 3D printing. The advantages
of 3D printing are that complexity can be added at no cost, no tooling is required, waste is reduced,
less operator skill is required, and some assemblies can be printed already put together. These
advantages allow for design alternatives to be printed and tested relatively quickly and at low cost.
Innovative, creative design concepts can be modeled and printed to test if the concept works. If it
does not work, the amount of time and resources lost are comparatively small to other design
processes.
3D printing also has some disadvantages. It has a limited and relatively expensive
selection of materials, is difficult to scale, results in parts with reduced mechanical properties, and
requires post processing. The printed parts must either be printed with a flat surface or with
additional raft or support material. The additional material takes time to remove and is wasteful.
3D printers are available for both personal and professional use. The cost of 3D printers
depends on their resolution and material usage, and can range from approximately $100 to
millions of dollars. The 3D printers used for this project to fabricate our designs were a XYZ Da
Vinci 1.0, Dimension SST 1200es, Makerbot Replicator 2, and Sindoh 3D printer. The
components were printed with PLA of a diameter of 1.77 mm. The printer used is a single extruder
and heats to 230 degrees Celsius that prints layers of .2mm. The components are all printed with
a 7% to 8% hexagonal infill on a non-heated glass bed.
The two most common materials for 3D printing are ABS and PLA. ABS, or AcrylonitrileButadiene-Styrene, is a plastic commonly used in 3D Printing. It is commonly used because of its
melting point temperature which coincides with safe heating practices in 3D Printers, as well as it
strong mechanical properties. It is also a relatively inexpensive and widely produced material so
it is beneficial for consumers to utilize it. PLA, or Polylactic Acid, is a thermoplastic typically made
from corn starch. Due to being made from natural materials, it is biodegradable. PLA is also
mechanically stronger than ABS so it is a viable candidate for 3D Printing. Similar to ABS, it also
has a melting point consistent with the temperature ranges of most commercial and industrial 3D
printers.
PLA was chosen as the material of choice for this project for two reasons. First, the fact
that it is stronger than ABS gives it an advantage for the project’s purposes. Second, is that it is
biodegradable. Wanting to sidestep the reality of leaving an environmental footprint with this
project, the team decided to pursue PLA as their material of choice.
22
2.4 Current Technology
There are currently many transradial prostheses commercially available and even more
under development. A few relatively inexpensive and innovative designs were evaluated and
analyzed for their strengths and weaknesses. These strengths directly influenced the prototype
design while the weaknesses were mitigated or eliminated.
2.4.1 Open Bionics’ Ada Hand
The Ada Hand is a robotic hand designed by Open Bionics. It is 5 degree-of-freedom, fully
articulated hand actuated by five Firgelli linear actuators and braided thread. All of the actuators
are housed in the palm of the hand. Open Bionics
describes it as “...perfect for anyone that is doing a project
with robotic hands or wants a neat, light, and functional
robotic hand for use with a humanoid robot. It's an
excellent platform for research into prosthetics, object
grasping
and
many
human-robot-interaction
applications”.33
Therefore,
despite
its
human
characteristics, it is not recommended or intended as a
prosthesis. However, due to its articulate nature, it can be
Figure 17: The Ada Hand32
used as inspiration for designs. Its very anthropomorphic
aesthetic and its ability to achieve numerous gestures are incredibly ideal for prosthetic hands.
2.4.2 e-NABLE’s Limbitless Arm
e-NABLE is an organization dedicated "...to [creating] free 3D printed hands for children
all over the world who have been born missing fingers or who have lost them due to accident,
illness or war”.35 The organization has produced over 2000 designs for those in need of a
prosthetic arm. Most of the designs are mechanically powered by a functional wrist or elbow. The
Limbitless Arm is e-NABLE’s first myoelectric design. “This experimental design was created for
individuals with above elbow limb differences by a team of students at University of Central
Florida”.36 The arm is an open-source design featuring an Arduino Micro microcontroller, a single
servo capable of producing torque around 12.1 kg-cm, muscle sensors, and Kevlar survival cord
to move the fingers. A disadvantage is that since only one servo is used, the hand can only open
and close. However, the Limbitless Arm acts is a decent baseline.
Ada Hand Kit, n.d.
Ibid
34 The Limbitless Arm, n.d.
35 Ibid
36 Ibid
32
33
Figure 18: Limbitless Arm34
23
2.4.3 InMoov
InMoov is an “...open source, 3D printed, life size
[humanoid] robot”.38 The arm features five servos with a servo
controlling each finger. The servos are controlled with an Arduino
microcontroller. Several videos of the arm have been posted
showing the hand switching between a wide array of gestures,
such as those for rock-paper-scissors, the “rock on” hand symbol,
the peace symbol, pointing, and a “count down” by closing fingers
individually.39 InMoov was designed for a specific robot and is not
a prosthesis. However, it is very usable as inspiration for
designing a hand capable of numerous ADLs and gestures.
2.4.4 Rehabilitative Robotic Glove
Although it is not a replacement for missing arms or
hands, the Rehabilitative Robotic Glove does act as an example
of durable string-based actuation. The Rehabilitative Robotic
Glove is a medical device designed by a WPI MQP meant to aid
Figure 19: InMoov Hand37
the rehabilitation of those who had recently undergone a stroke.
“This glove utilizes a cable system to open and close a patient’s
hand. The cables are actuated by
servomotors…”.41 The system uses Kevlar
k49 cables due to their tensile strength and
low elasticity. Cables are placed on each
side of each finger and attached to a
respective servo and spool. This, as a
result, always pulls the fingers to a desired
position based on which direction the
servos move. This type of system could
replace elastic thread used in traditional
prosthetic designs. The benefit of this is the
lack of elongation of the thread that can
effectively render the hand useless after a
period of time. However, more cables
Figure 20: Rehabilitative Robotic Glove40
would need to be routed throughout the
hand and proper sizing of the spools would
be required.
DivideWorks, n.d.
Ibid
39 Ibid
40 Delph et al., 2012
41 Ibid
37
38
24
2.4.5 Multi-DOF Anthropomorphic Prosthetic Hand
The Multi-Degrees-of-Freedom Anthropomorphic Prosthetic Hand developed at the
Harbin Institute of Technology’s Robotics Research Institute has some of the most sophisticated
hardware and software considering its size and anthropomorphic design. The thumb, the index
finger, and the other three fingers are each actuated by a DC motor respectively that are all
contained within the palm. The fingers operate on an underactuated mechanism that allows for a
wide variety of self-adapting grips, and an opposable thumb actuates on a spherical bearing
allowing for a force and pinch grips.
The hand enclosure is made of aluminum,
and the components comprising the mechanical
linkages have been fabricated out of steel. This
means a redesign of the linkage would be
necessary for an additive manufacturing process.
However, what separates this hand from other
designs are the sensors it utilizes to implement a
control system for the position of each finger. A
torque, position, and force sensor is included in
each finger and a low-power microcontroller uses
the input from these sensors to drive the motors.
Figure 21: Three motor design for anthropomorphic
hand
By implementing a control system like this, it is not
necessary to pre-program grips as the fingers will
automatically adjust to the object being grasped.42
Figure 22: Control System for Force Feedback Controlled Finger
42
Zhao, 2006
25
2.4.6 Design of a Human Hand Prosthesis
This Major Qualifying Project from WPI in 2012 gives additional insight into the
implementation of a mechanical linkage design in a prosthetic. The “...design incorporates five
individually actuated fingers in addition to powered thumb roll articulation...” and also has a simple
feedback control system in the form of a force sensor and light emitting diode (LED) built right
onto the forefinger. The actuation system for the four fingers is very limited consisting of a single
bar connecting the knuckle joint to the distal phalynx. The linkage could be easily converted to a
3D printable design and then modified for additional degrees of freedom.
The mechanical linkage for the thumb attempted to maintain the two degrees-of-freedom
of a human hand by using a two motor system for pitch and roll respectively. The motors
controlling both axis can be controlled independently adding three additional grip capabilities.
Each finger including the thumb has a potentiometer mounted to each rotational axis which gives
the microcontroller an idea of phalange positioning for further motor control.43
Figure 24: Finger Actuation System
Figure 23: Mechanical Linkage for Finger Actuation
43
Ventimiglia, 2012
26
2.4.7 Socket
Socket design is crucial to the success of the
overall prosthesis. Well thought out socket designs
and careful consideration of residual limb
preservation set the stage for patient success-maximizing range of motion, providing stability
throughout daily activities, and comfortably
distributing the forces exerted on the residual limb
during movement and suspension. In contrast, poor
socket design will often drive a person to abandon
the prosthesis.44 The challenge of achieving the best
socket fit is that the residual limb changes shape and
size overtime.45 Multiple design approaches have
evolved to achieve this. There are five standard
design approaches: vacuum suspension, bladdercontrolled, compression/release stabilized (CRS),
3D printed, and mechanically adjusted sockets.
2.4.7.1 Current Socket Technology
Sockets that don’t fit and are uncomfortable
lead to critical health problems. The health problems Figure 25: Photos of residual limbs suffering from
skin problems
can be severe enough that the added benefits the
functions provide are outweighed by the additional health problems. When this happens, the
prosthesis is useless and unused. Sockets that don’t fit can cause skin, vascular, and lymphatic
problems, tissue damage, nerve damage, pain, and discomfort. All types of prosthetic sockets
can cause skin problems. Unfortunately, these problems are often dismissed, forgotten, or
ignored.
The cause of these skin problems is twofold. First, these problems are caused by
increased normal, shear, and frictional forces on the residual limb. Secondarily, they are caused
by layers of socks or gel sleeves surrounding the residual limb.46 The socks are a necessary
component of many types of prosthetic sockets in order to achieve a tight fit. It is common to wear
layers of thick socks to compensate for daily changes in the size and volume of the limb.
Unfortunately, the socks and the socket itself insulate the residual limb, building up heat. This
residual heat causes sweat on the skin. The sweat compounded by a lack of airflow leads to a
multitude of skin problems.47 Additionally, sweat contributes to prosthetic odor, which was
identified as a problem in the survey of amputees.48 These skin problems include but are not
limited to: ingrown hair, rashes, skin irritation, odor, erythema, blisters, ulcers, and skin thickening.
Lake, 2008
Ibid
46 Maguire & Boldt, 2013
47 Mak, Zhang, & Boone, 2001
48 Legro et al., 1999
44
45
27
The vascular response of the residual limb is also important because ischemic injury, the
restriction of blood flow to the limb, can cause pressure sores and localized malnutrition.
Epidermal forces, forces against the skin, are a major factor associated with vascular problems.
More explicitly, decreased blood flow occurs with increased application of either normal or shear
forces. In addition, the prosthetic device affects the lymphatic system. Lymphatic function is
associated with skin health in the form of tissue edema.49 Accumulation of lymphatic waste could
occur if external forces hindered the flow of lymph fluids, a situation which could occur from the
forces applied to the limb in a socket.
The fit of the socket can address these issues. The fit of a prosthetic socket is the most
important feature as identified by amputees in a survey when compared to other factors such as
weight or ease of use of the prosthesis. The fit of the socket is the key to maximizing range of
motion, providing stability throughout daily activities, and comfortably distributing the forces
exerted on the residual limb. The challenge of achieving the best socket fit is that the residual
limb changes shape and size overtime.50 Over the course of months, the residual limb can
drastically change in volume with the loss of muscle mass. Over the course of hours and days,
the residual limb has smaller volume changes.
One research team, concerned about the effects of slippage, due to improperly fitted
sockets, in sockets causing unnecessary frictional force, measured the amount of slippage that
could occur for sockets to retain a good fit. Well-fitting sockets had slippage of 2mm to 6mm.51 In
this study, researchers concluded slippage substantially greater than 6mm causes user distrust
of the prosthetic limb and severe friction on the limb causing the aforementioned frictional skin
problems. Slippage less than 6mm puts extra pressure on the limb causing issues such as
pressure sores and ulcers. Intensity and duration of load application of the residual limb are
inversely related to ulcer production. Additionally, this increases limb temperature, contributing to
sweat production in the socket.52 Many types of skin problems such as these can deter amputees
from using the device as even mild skin problems cause discomfort and can lead to infection or
ulcers if not treated correctly. It is estimated that 75% of amputees will experience skin issues,
causing possible decrease in socket use.53
The way the socket manages and adjusts to these changes in limb size and volume
directly impacts the fit and comfort of the socket. There are multiple methods that sockets adjust
to these changes. The most common methods are: vacuum suspension, bladder-controlled,
compression/release stabilized, exact limb replica, and mechanically adjusted sockets.
2.4.7.2 Vacuum Suspension Sockets
Vacuum suspension is a socket method for managing residual limb volume. The user
wears a liner or prosthetic sock that forms an airtight seal inside the socket, as shown in Figure
26. Skin suction is a type of vacuum suspension that incorporates a one-way valve to create a
seal between the limb and the socket. This seal exerts a positive pressure when inactive and a
Mak, Zhang, & Boone, 2001
Lake, 2008
51 Commean, Smith, & Vannier, 1997
52 Mak, Zhang, & Boone, 2001
53 Highsmith & Highsmith, 2007
49
50
28
negative pressure when active. Other forms of vacuum suspension use a pump to remove air
from between the limb and the socket which applies a constant negative pressure to the limb.54
Vacuum suspension has the added benefit of managing residual limb volume and can increase
total limb volume to better keep the skin in contact with the liner and socket.55 The technology
also maintains a better socket fit than standard prosthetic sockets, reducing discomfort or pain
from wearing the prosthesis throughout the day. However, these types of sockets need to initially
be fitted to the individual for effectiveness. Also, for larger volume changes, the addition of socks
or small air bladders in the socket is required. Furthermore, they can cause skin issues due to
sweat accumulation or an allergic reaction to the socket material. This is because the gel liners
used can cause ingrown hairs or rashes.
2.4.7.3 Bladder-Controlled Socket
Bladder-controlled sockets adjust their fit by increasing the pressure of one or more
bladders on the inside of the socket. The bladders apply pressure to the residual limb to hold it in
place. The bladders accomplish this with either pneumatic or hydraulic systems that inflate their
bladders with a fluid. Most bladders are manually pumped up or deflated with ambient air to their
desired comfort and pressure level. An advantage to bladders is that they can be added to an
existing socket and filled with air or liquid in order to replace volume lost by the residual limb, as
shown in Figure 26.58 The Smart Variable Geometry (SVG) socket uses liquid-filled bladders
powered from the cyclic motion of walking.59 The liquid-filled bladders replace volume lost by the
user's residual limb and apply pressure to the limb to keep the user's limb in place. The device's
maximum pressure can be adjusted by a prosthetist who also decides on the number and
placement of bladders. The device's main advantage is that it does not rely on electrical power to
pump up the bladders. However, this means that the user must be in motion in order to pressurize
the bladders. Small motions can potentially causing a loose fit and irritation of the residual limb.
The SVG design also does not allow for
the bladders to be individually controlled.
Individual control of the bladders could
allow for more precise control of volume
loss in the socket and increased comfort
for the user. A disadvantage is that higher
pressure can have negative effects on the
user's limb, potentially cutting off blood
flow and causing tissue damage.60 The
controls and fluid are also bulky and
heavy.
Beil, Street, & Convey, 2002
Street, 2007
56 Hedef, n.d.
57 Sanders, 2001
58 Sanders & Cassisi, 2001
59 Greenwald, 2003
60 Sanders, 2001
54
55
Figure 26: Vacuum suspension socket56 and Inflatable
bladders57
29
2.4.7.4 Compression/Release Stabilized Socket
The Compression/Release Stabilized
(CRS) socket uses pre-compression to save
amputees’ energy. Conventional prosthetic
sockets only compress the residual limb during
movement. When an amputee with a
conventional prosthetic socket wants to move
their limb, they must first compress the soft tissue
between the bone and the socket before any
movement occurs. This compression is done
subconsciously as part of regular movement with
a conventional prosthetic socket. This extra
movement wastes the time and energy of the
wearer. Due to the pre-compression of the CRS
socket, all energy output directly moves the limb.
This leads to increased energy output efficiency,
range of motion of the end effector of the limb,
and user control of his or her limb.62
The CRS socket raises safety concerns Figure 27: Compression/Release Stabilized Sockets61
such as blood flow restriction. While the total area
of compression is reduced, there is higher compression placed on the remaining areas of the
limb. This high compression could cause further issues for those with blood circulation problems,
such as those with heart disease or diabetes. Another concern is discoloration of the limb, which
can indicate soft tissue damage. Once the CRS socket is removed after wearing for 3-4 hours, it
can take another 3-4 hours for the reduction of redness in the residual limb. This discoloration
also brings up concerns about long term use and if the device would cause a deformation of the
limb.63 While there are currently no user reviews about the CRS socket because the device is still
in clinical testing, an overall review of prosthetic users who used restrictive sockets similar to the
CRS socket was analyzed. Users reported that for the first 2-3 years, they felt discomfort at the
areas of restriction.64
Bio Designs, n.d.
Alley et al., 2011
63 Ibid
64 Legro et al., 1999
61
62
30
2.4.7.5 Exact Limb Replica
The importance of a socket’s fit led to exact limb replicas. Initially molds and casts were
made of the limb to create sockets that fit snuggly. Now, advances in scanning technology and
3D printing, have led to multiple amiteur prosthetic and socket designs. Every residual limb is
different and custom sockets must often be individually fabricated. 3D printing simplifies the
process of custom sockets. The key elements of this process are the measured data and the
recorded residual limb topology. It is essential for this data to be both precise and accurate for
the socket to fit correctly. The advantages of this process is that the prosthetic socket is an exact
fit for the amputee’s residual limb. Unfortunately, the process often requires several prints and
fittings before a sufficient level of comfort is achieved.66 The socket also doesn’t account for
volume fluctuations in limb size. The amputee still has to wear multiple socks in order to adjust
for any day-to-day fluctuations, causing heat retention, chaffing, and rashes.
Figure 28: 3D Printed Socket Creation Process65
2.4.7.6 Mechanically Adjusted Socket
Mechanically adjusted sockets are often static systems with multiple manual controls that
make micro-adjustments. These controls allow the user to make small adjustments throughout
the day to accommodate change in limb volume. Larger changes can also be made as the limb
changes over the course of months or years. Revolimb, shown in Figure 29, powered by Click
Medical’s Boa Closure System and Martin Bionic’s Socket-less Socket, also shown in Figure 29,
are examples of mechanically adjusted sockets. These sockets allow the user to easily and
quickly adjust the fit of their socket to their current needs. The user can adjust for a tighter fit for
more performance, or reduced compression for resting and sitting. The Revolimb system utilizes
a series of panels along the outside of the socket that add or reduce compression to the limb with
the turn of a dial.67 The Socket-less Socket utilizes all adjustable carbon struts as a framework
with flexible cross connectors and dynamic straps so that the entire socket is adjustable. For daily
adjustments, the straps easily ratchet to adjust compression. For long term limb changes, the
Devadass et al., n.d.
Ibid
67 Click Medical, n.d.
65
66
31
carbon struts and cross connectors fastened by rivets can change the overall shape and size of
the prosthetic socket.68
Figure 29: The Revolimb69 and the Socket-less Socket70
2.5 Impacts and Ramifications
2.5.1 Manufacturing, Sustainability and Environmental Aspects
To manufacture the prosthesis the current process would take time. Each prosthesis
would have to be hand assembled and all the parts would have to be gathered separately. The
current process for creating the hand is to use 3D printed technology. The amount of time it would
take to print every piece in mass quantity would not be worth the time and effort. However, once
a patient has a prosthesis it should last for an extended period (at least 6 months). There may be
a need to have replacement parts if something were to break and having the parts 3D printed is
advantageous. The model could be sent to a 3D printer close to the patient and they could even
pick it up the same day if the printer is not being used. If the patient has their own 3D printer then
they could print a new part and immediately have it. Since most of the parts of the prosthesis are
3D printed they could keep using it as long as the electronics work. Should the electronics fail,
after ordering new parts they should be able to get them within 3-5 days and continue to use the
prosthesis.
The material that the prosthesis was printed with is Polylactic Acid or PLA. It is a
biodegradable substance that usually takes 30 days to be completely dissolved. Compared to
other devices on the market that use other materials like plastics or metals when the device is
discarded it is not recycled. Other devices cannot be broken down and have to be thrown away
when they are finished being used or break.
Martin Bionics, n.d.
Click Medical, n.d.
70 Martin Bionics, n.d.
71 Baily, n.d.
68
69
32
The components cost 375$ to obtain. With the addition of using a service to print the
components it could roughly cost 450-500$. For prosthetics with the functionality of the one
designed it would cost hundreds if not thousands of more dollars to achieve the same functions.
71
Using 3D components makes it less expensive and easier to manufacture than other models
on the market.
2.5.2 Ethical Concern and Health and Safety Issues
The highest design specification of the prosthesis was that it would be safe to use over
long periods. The biggest area of concern would have been between the socket and the patient
as that is where the prosthesis attaches. The skin could grow irritated over time if the wrong
material was used for padding. From testing, it showed that there was minimal slippage and
irritation on the skin after the contact points of the socket was worn. Other aspects of the
prosthesis should not cause any harm to the patient while they are using it. The prosthesis was
designed to help those in need and provide them with a way to try to return to normalcy.
Since there are many people suffering from the loss of a limb it is difficult to provide a way
to help everyone. Those who served in the military, for example, should they receive a prosthetic
for free for helping serve the country? There is a dilemma currently that the products on the market
to restore functionality and normalcy are very costly. It is not feasible to provide a free prosthetic
to everyone who needs them, but at the same time it is very hard to decide who does receive one.
The current design could fill the void that currently exists. Not everyone is able to afford a fully
functioning prosthetic and the prosthesis designed could change that. More people would be able
to afford it and would not have to make tough decisions of who gets assistance and who does
not.
2.5.3 Economics, Political, and Societal Impacts
There are many prosthetics on the market today. The cost could be between a couple
hundred dollars to even a thousand dollars. People who have gone through traumatic experiences
who have lost a limb may not be looking for one that is able to do everything their old limb could
do. They may want to have a prosthesis that can do the basics and help them get by. The less
expensive prosthetics could provide the bare minimum needs, but may need constant fixes or
may be inconvenient to use. More expensive prosthetics could provide all the functions desired,
but is generally too expensive to purchase and maintain over time. Our current 3D printed
prosthetic device would allow a much lower cost and offer the same functionality as the more
expensive ones.
If the designed prosthesis were to become a commercial product it would most likely
change the market of prosthetics. The amount of people that would be able to obtain a prosthetic
would increase and if in the future there are more and more people in need it could continue to
help everyone, not just the people who can afford an expensive model. Those who might not have
the health insurance to cover the cost of a prosthetic would be able to obtain one as well. It would
help not just those effected now, but even people in the future who might need assistance later
in life.
33
3. Methodology
Standard methods for engineering design were utilized. The problem was articulated in
detail. Design goals were generated, defined, and ranked to describe an idealized design. The
terms of a successful solution were defined by the design specifications. The ideation process
generated multiple possible designs. These design alternatives were evaluated and ranked
against the design goals and specifications using decision matrices. The highest ranking and best
solution was selected to move forward. The selected design was modeled and a prototype
manufactured. Tests were created based upon the design specifications. The prototype
underwent these tests to determine if the design met the design specifications and was
successful. The tests results show the strengths and weaknesses of the design, revealing areas
for further development.
3.1 Design Goals
The project design goals were identified, defined, compared and contrasted, and ranked.
The design goals identified were: safety, ease of use, durability, comfort, range of motion,
maximum capable load, serviceability, weight, scalability, cost, ease of assembly, environmental
factors, aesthetics, time to assemble, and time to print. The design goals were compared against
each other using a weighted decision matrix. If the vertical design goal was more important, the
value “1” was recorded. If it was of equal importance, the value “0.5” was recorded. If it was less
important, the value “0” was recorded. The total was then tallied. The design goal with the highest
value was ranked as number “1” and the least as number “15” and so on. The weighted decision
matrix used to determine the ranks can be found in Appendix D.
Ranked Design Goals
1. Safety
2. Ease of Use
3. Durability
4. Comfort
5. Range of Motion
6. Maximum Capable Load
7. Serviceability
8. Weight
9. Scalability
10. Cost
11. Ease of Assembly
12. Environmental Factors
13. Aesthetics
14. Time to Assemble
15. Time to Print
34
Safety
The safety of the prosthetic is imperative due to the fact that it is meant to assist a patient
in their daily lives and not cause them harm. If the device is not safe, then it becomes effectively
useless and will not be used. Safety includes many factors such as excessive heat, electrical
shocks, reduction of blood flow, ingrown hair, rashes, skin irritation, odor, erythema, blisters,
ulcers, and skin thickening
Ease of Use
Ease of use is an incredibly important factor to consider when designing a device. If the
user cannot understand how to use or does not have the abilities needed to use the prosthetic,
then there would be little reason to use it. The controls must be easy to understand and use simple
functions for anyone to be able to use the prosthetic.
Durability
The durability of a prosthetic is crucial. It should be able to withstand the daily loads it is
put under with little wear. A device that cannot withstand these loads or fails after a short time
and constantly needs repair presents itself as more of a hindrance than an asset. The more
durable the prosthesis, the less it would need to be serviced, reducing both the cost of the
maintenance and the time the user cannot use their prosthesis.
Comfort
Many different factors are considered for the comfort of the prosthesis. The socket where
the user and the prosthesis are joined must be comfortable over long periods of time. The amount
of heat retained when wearing the prosthetic must be low so it can be worn over long periods of
time. It must cushion the amputation site so it does not irritate the skin and it should be secure
without causing stress on the existing arm.
Range of Motion
The prosthetic must complete a wide range of motion to ensure the ability to aid in daily
life. It must have the capability to complete a number of gestures to allow a person can continue
to live a normal life. If the hand does not have a wide range of motion, then there would be no
point in using the prosthetic for daily use.
Maximum Capable Load
The amount of weight that the prosthetic can successfully hold is another important factor.
If it cannot support a reasonable amount of weight, such as a gallon of milk (8.6 lb.), it would not
be able to complete many of the activities in life. A prosthetic incapable of handling these weights
serves only the purpose of aesthetics.
Serviceability
The easier the prosthetic is to service, the more desirable it becomes for potential users.
With expected wear-and-tear, if a part breaks or fails over time, but can be fixed within a relatively
short amount of time and with little effort, that prosthetic becomes far more desirable to that given
user.
35
Weight
The weight of the prosthetic is important because if it was attached to a patient and it could
not be lifted by them, there would be little use to it. The prosthetic must be light enough for people
of different builds to be able to use it for extended periods of time without feeling tired or
inconvenienced. With the inclusion of many different electrical components, the weight needs to
be focused on or the device could easily become too heavy and unusable.
Scalability
There are different kinds of people that need prosthetics and this project aims to cater to
them. If there is a need to change the sizing to make a better fit for a patient, it is important to be
able to do so to allow more people to benefit from the prosthetic.
Cost
With different components being used in the prosthetic, it must be taken into consideration
who will be able to afford the device. The cost must be kept reasonably low while maintaining a
high standard of quality so there is the possibility for more people to be able to use the prosthetic.
Ease of Assembly
When a patient obtains the prosthetic, they would need to be able to assemble it
themselves with the instructions provided. It should be intuitive and simple to follow the
instructions so patients are able to do it themselves. Making a prosthetic that worked, but needs
others or a professional to assemble would not be practical or useful to a patient that has no
access to either.
Environmental Factors
The hand itself must be able to withstand the constant use throughout different
environments. It cannot fail because the patient uses it in the sun or splashes water onto it. The
purpose of the prosthetic is to allow the patient to continue in their daily life and having to
significantly limit them to avoiding certain conditions.
Aesthetics
A key purpose in having a prosthetic is the desire to return to normalcy. If the hand was
aesthetically unpleasing or did not resemble a human hand, it would not be desired by many. It
should function well while at the same time being a product that people would willingly to use and
show.
Time to Assemble
When considering the patients using the prosthetic, it is necessary to think about how long
it will take them to assemble it. If a large amount of time is required to put the prosthetic together,
it could frustrate many and deter them from using the product. It should be able to be put together
in a relatively short amount of time with as little stress as possible.
Time to Print
The prosthetic should not take an extensive amount of time to print if there was a need to
recreate a certain part. 3D printing is not instantaneous. As such, a part or component should be
36
designed to take as little printing time as possible as to allow for the part to be efficiently replaced
or fixed.
3.2 Design Specifications
The prosthetic will be designed to fulfill the following specifications in Table 4. These
specifications were determined through by combining the weighted design goals and the design
decisions previously discussed.
Table 3: Design Specifications
Specification
Quantitative Value
Qualitative Description
Safety
N/A
Will not harm the user through scratching, cutting,
burning, or be able to complete involuntary movement.
Maximum
Operating
Temperature
140°F (60°C) +/- 10°F
(5°C)
At around 140°F, PLA will reach its glass transition
temperature
Minimum
Operating
Temperature
-10°F (-23.3°C) +/10°F (5°C)
At around -10°F, most rubbers will begin to reach their
glass transition temperature. Significant performance
degradation will occur around 0F.
Ease of Use
N/A
Will be easy for users to trigger movement with simple
movements. Will not feel awkward or be causing unusual
or unnatural movements to trigger function.
Life Expectancy
2000 cycles, with
components being
replaced when
needed.
The base of the prosthetic will be able to last while
components that wear will need to be replaced.
Additionally, number of times the hand can open and
close before reduction in performance will be weighed
into this goal.
Socket Fit
Comfort
minimum of 5+ on a 1
to 10 scale
Socket comfort score (SCS) based upon numerical rating
scale (NRS), 0 being least and 10 the most comfortable
score71
Range of Motion
Wrist
0 to 180 +/- 10
degrees, 1DOF
Angle 0 when resting the palm flat to resting the back of
the hand flat
Range of Motion
MCP
0 to 90 +/- 5 degrees,
1DOF
Joint and knuckle. Angle 0 at full extension, curving
inwards
71
Hanspal, Fisher, & Nieveen, 2003
37
Range of Motion
PIP
0 to 90 +/- 5 degrees,
1DOF
Joint at end of proximal phalanx
Range of Motion
DIP
0 to 90 +/- 5 degrees,
1DOF
Joint at end of intermedial phalanx
Range of Motion
TMC
Opposable thumb,
2DOF
Joint at base of thumb
Range of Motion
IP
0 to 90 +/- 5 degrees,
1DOF
Joint at distal end of thumb
Maximum
Capable Load
5 to 11 lbf (22-49 N)
+/- 1 lbf (4.5 N)
Total allowed force on the fingers from carrying a weight
Serviceability
N/A
Universal replacements will be readily available.
Information for purchasing components will be provided
with the product.
Weight
5 lb. (2.27 kg)
Weight of the entire assembled prosthetic and socket
Scalability
Provide dimensions
and sizes for ages of
16 and up.
Based upon average size of fingers and capable grip
strength. Equations and relations in the model will be
created so more customers can benefit from the product.
Cost
$1000 +/- $100
Cost of all printed and non-printed parts, including wires,
electronics, hardware, etc.
Difficulty to
Assemble
16 year age minimum
for assembling the
product.
16 years is a common age for safety understanding. At
this age, the person will also be able to fully understand
written directions that will come with the product.
Environment
Factors
Resist water from
electrical components
The product will resist water from the electronic.
Aesthetics
6+
Numeric scale of 0-10, 0 being least and 10 being most
appealing.
Time to Assemble
6 +/- 1 hours
Cleaning/finishing and assembling parts and electronics
with the given tools by someone 16+.
Time to Print
<60 hours
Time for all parts to print.
38
4. Design
With sufficient research completed on all aspects of a prosthetic, and design goals and
specifications established, designs were developed for each component of the proposed
prosthetic. The designs include models developed in SolidWorks that are able to be fabricated
using additive manufacturing along with a description of all materials necessary to construct the
component. Additionally, diagrams and flowcharts are provided to describe the operation of
electrical systems of the design. The following sections present the designs for each component
of the prosthetic hand. For ease of discussion, the diagrams in Appendix G: Model Prototype
Images will be used to denote which portion of the prosthetic is being referred to.
4.1 Hand
The hand must be durable, functional, and have a sense of natural movement to the user.
The hand has been designed with three elements: the palm, the thumb, and the fingers. The palm
in this project will serve the purpose of guiding the Kevlar thread to its appropriate finger and
housing electrical components for finger movement. The thumb is its own element due to its
important role in synchronizing with the others fingers for gripping purposes. The main difference
will be in its physical design while its movement method and mechanic system will be the same
as the fingers. The thumb will serve mainly as
support and assistance in gripping objects
and supplying the force necessary for
interaction with various objects. The fingers
play the largest role in the hand and thus will
be three-jointed. This adds a sense of reality
and functionality for the user and also aids in
gripping and interaction with objects. The
image below reflects the final model. The final
model utilizes 3D printed pins for the joints.
4.1.1 Finger Design
The fingers were the primary focus of
the hand as they will have the most
interaction between the prosthetic and the
objects being handled. The hand must be
able to complete all essential ADL’s and
therefore must have some system for finger
actuation. Actuation here refers to the system
by which the fingers will be set in motion,
enabling them to open and close. Through
research the team found multiple methods for
achieving such actuation, but only a few
Figure 30: SolidWorks Model of Hand Design
39
allowed the range of motion necessary for significant ADL completion. The
design being pursued involves using Kevlar threads to both pull the finger
into its closed position and then back into its open position. This method
gives the fingers the ability to adapt to object shape and size and also
provide the necessary force required to perform individual, everyday tasks.
The actuation of the fingers is explained below.
The fingers each contain channels that have threads through them
to actuate the fingers. These pulley channels are referred to as “close” and
“open” and run along the bottom and top tracks respectively. Both cables
reside on one side of the pin connecting each finger segment. When
activated, they provide a moment on the pin causing the segment to move
towards the direction of the cable. Activation occurs via servo motors. The
thumb, index finger, and middle finger each have a dedicated servo. The
remaining two fingers use a single servo motor, resulting in a total of four
servos used in this design. Each of these motors will have their standard
parallel gear replaced with a two piece track which can be seen below. The
only exception to this design is the servo actuating the pinky and ring finger
will have a four piece track to actuate two fingers simultaneously.
When the thread is pulled by the servos, the threads in the close
channel will torque around the pins that will cause the finger to actuate to
its closed positions depicted by Figure 32
The finger’s joints will actuate until the object it is gripping forces it Figure 31: Push Pull
to stop. The other joints will actuate until the finger is completely gripping
Pulley Design
the object with the necessary force.
Each piece of the track will be dedicated to either the “open” or “close” cable. These cables
will be fixed via screw to one tangential location on
the track. The process will be identical for each
cable; the only difference being that each cable will
run in opposite directions. When the servo is
activated, for example to close the hand, the “close”
cable will begin wrapping around the track until the
servo is deactivated. The same process would
occur for the “open” cable to open the hand when
the servo was activated in the opposite direction.
Having both cables is essential to this design as it
results in fewer moving parts, removing the need for
Figure 32: Fully Actuated Finger
springs, and the ability to use wear-resistant
materials.
40
4.1.2 Palm Design
The palm’s purpose in this assembly is to
channel the finger actuation threads and act as a
central connection point for the fingers. All the
fingers will connect to the palm with 3D Printed
joint pins, much like the ones in the finger, but with
a longer length. The threads will be run through
channels within the palm to their desired
destination.
The channels have been designed to not
interfere with one another or cause a great
amount of friction to dampen the actuation of the
fingers. All angles have been designed to be 44
degree or less as not to subject the thread or palm
to any unnecessary torque or grinding.
The electronics that are placed in the hand
are contained in an assembly called the Palm
Cover, pictured below.
The Palm Cover is comprised of a bed and a
cover. It will house the battery and the control board
that runs the hand. The bed is placed over the back
of the palm to divide the Kevlar thread and the
electronics. The battery and control board will then be
placed on the bed, and then covered by the cover.
Two 1.5” screws will be screwed through the cover,
bed, and the palm to ensure it will be enclosed.
Figure 33: Palm Design
4.2 Forearm
The forearm in the scope of this project will
range from the end of the socket to where the hand is
connected. This area does resemble a human
forearm and has allowed for simplicity in definitions
and comparison between amputees and nonFigure 34: Palm Cover
amputated individuals. The forearm will serve the
main purpose of housing the mechanics of the hand
including servos and their rail system for Kevlar string actuation. These essential components will
be housed within an enclosure to preserve the functionality of the prosthesis.
41
Figure 35: Render of Full Assembly
The servo frame housing has been
designed for longevity and robustness. In the
design, screw holes are 0.164 inches in
diameter to properly fit #8 bolts in this assembly.
The servo frame to palm connection will be
completed with two bolts tucked inside the servo
frame providing the design with stability and a
low overall length. The servo frame allows for all
four desired finger actuating servos to push and
pull cables which travel from the servo rails to
the tips of the fingers. These rails, known as
cable guides, are fixed to the servo heads and
have a 42mm diameter. This size was selected,
because it is the smallest diameter which still
Figure 36: Servo Frame, without Servos
allows the cable to perform a full actuation with
the HS-5585MH motor. The servo housing
frame was also specifically designed around these servos, and its main cavity is exactly the size
of four HS-5585MH servos positioned side by side. To fit four 42mm cable tracks, the servos
alternate sides (left and right) and orientation (up and down). In addition to this, the disks are
staggered, the rear two servos sit further out from the center of the frame than the front two.
Combining all of these space saving techniques produces what we believe to be the smallest
possible forearm for our prosthesis The servo frame mates with the socket and palm via two #6
screws for each connection.
42
Table 4: Socket Frame Decision Matrix
Note: the design alternatives are not comparable in every category.
4.3 Socket
The socket design alternatives were rated based upon their fulfillment of the project design
goals and specifications. Once they were rated and totaled, the highest rated design alternative
was selected. The design alternative was benchmarked by existing similar products. Those
products were evaluated for possible improvements. The final design concept was generated and
a prototype fabricated.
The five alternative socket frame design categories were: vacuum suspension, bladdercontrolled, compression/release stabilized (CRS), 3D printed, and mechanically adjusted sockets.
These designs were evaluated against the ranked design goals in Table 5. It was decided that
the socket frame will be mechanically adjusted and the structural components 3D printed.
The individual parts of the mechanical socket were identified and designed. Figure 38
shows the full assembly of the socket frame. The socket frame parts are: 1 ‘base plate’ (red), 3
‘base plate-hinge connectors’ (yellow), 3 ‘hinges’ (green), 3 ‘struts’ (blue), 3 ‘strap retainers’ (pink),
3 foam EVA pads, and 2 straps with buckles. These parts work together to provide an easily
adjustable, lightweight, sturdy, and comfortable socket frame. Forearms come in many different
shapes and sizes. The average forearm has an amorphous ovular cross-section that increases in
43
Figure 37: Socket Frame CAD Model
size from the wrist to the elbow. Muscular forearms have larger, more toned muscles that increase
the girth of the forearm. The change in girth from the wrist to the elbow is also more drastic.
The base plate and base plate-hinge connectors’ purpose is to adjust the position of the
struts for the girth of the forearm. The base plate has large and small sizes where the hinge-base
plate connectors can be attached. This allows adjustment for the girth of the forearm. There are
three arcs that comprise two sizes that can be individually altered for each of the three struts. The
base plate allows for the maximum adjustments for the individual’s comfort and needs. The
position of the struts can be swiveled within 90° and the diameter can be changed from 2 to 3
inches.
Figure 38: Base Plate-Hinge Connectors’ Positions for Girth Adjustments
44
The arcs also allow for the position and angle of
the struts to be changed to the individual’s comfort and
needs. The three struts are typically positioned
approximately 120° from each other. One struts lie on the
anterior (top) and two on the posterior (bottom) of the
forearm.
The hinges’ purpose is to adjust to the angle the
strut lies along the forearm. The hinge adjusts so that the
socket frame can fit all shapes and sizes. The majority of
forearms have a shallow angle and increase minimally in
girth when comparing the wrist to the upper forearm.
However, the drastic change in girth of muscular
forearms creates a steeper angle. The hinge allows the
socket frame to adjust to the individual's comfort and
needs. The hinge is adjustable from 90° to 180°.
Figure 39: Base Plate-Hinge Connector Angle
The struts’ purpose is to lie flush along the
Adjustment
residual limb. Three struts positioned around the residual
limb are the main structural support for the rest of the
prosthetic. They have an inward compressive force
against the limb. This force supports and offsets the
weight of the entire prosthetic. It is important for the struts
to have a large surface area so that the force is distributed
rather than being focused on one part of the residual limb.
However, the larger the surface area, the less breathable
the socket frame is. An encapsulated socket that
surrounds the residual limb completely provides the most
force distribution. However, it also has the highest heat
retention. It is necessary to find a balance between force
distribution and heat retention that is the most
comfortable. The struts range in size: the small is 4.5”,
Figure 40: Hinge Angle Adjustment
medium is 6”, and large is 8”.
The inside of the struts will be lined with supercushioning high-strength EVA foam. This type of foam is the same as what is used inside football
helmets. It is made for the outdoors and is ⅜ inches thick. EVA foam provides an added layer of
comfort.
3D printing with PLA is the best manufacturing process and material for the struts, base
plate, hinge, and base plate-hinge connectors. It is a relatively inexpensive material and quick
and easy to print more components. It takes 9 hours and 20 minutes to print all 10 of the socket
frame components. Once the amputee has learned how to fit, don, and doff a prosthesis,
understands their prosthetic preferences, and their residual limb is not swollen or changing shape
and size drastically, a more permanent socket frame can be constructed. The semi-permanent
frame replaces two of the struts, hinges, and base plate-hinge connectors with carbon fiber struts
at the preferred angle and proper girth. The third strut, hinge, and base plate-hinge connector are
45
not changed so that micro-adjustments can still be made. If
desired, the third strut can be made of carbon fiber and curved
so that it lies flush with the skin.
The socket frame utilizes fasteners and washers. The
socket frame 3D printed parts are held together by multiple
fasteners. A stainless steel binding post is used as the pin to
connect the hinge and hinge-base plate connector together.
Philips head screws and low-profile binding posts are used
together to allow easy adjustment to the parts. Nylon plastic
washers are located along the axes of the hinges to reduce
wear. Split lock washers are placed between the Base Plate
and Base Plate-Hinge Connector to reduce movement and
slippage.
The socket frame is held to the residual limb with ladder
straps and ratchet buckles. Ladder straps and ratchet buckles
Figure 41: Small, Medium, and Large
are typically used to secure rollerblades and snowboards.
Strut Lengths
Retainers help to hold the ladder straps in place along the struts
while allowing the straps to adjust freely.
An Under Armour compression sleeve or a prosthetic sock is worn under the socket frame.
The sleeve is relatively thin so as not to retain heat, provides compression to the residual limb to
reduce swelling, and reduces friction between the skin and the socket frame to prevent abrasions.
Velcro dots or “Shark Skin”, a product of Martin Bionics, can be added to increase friction between
the sleeve or sock to reduce slippage.
Figure 42: Non-Amputee Showing Strut Positioning
46
4.4 Motor Selection
A current design constraint for the hand is
for it to be capable of holding a jug of milk (around
10lbs/4.5kg) while maintaining the ability to
accomplish standard ADLs. The average human
finger is around 10cm (3.72in). As such, the torque
generated in a worst-case scenario of a 10lb milk
jug at the ends of the fingers of this length would
be around 595.2 oz-in (~42.8 kg-cm) (torque =
length of fingers x force). Furthermore, the torque
generated from the fingers themselves (assuming
PLA) would be around approximately 4.73oz-in
(0.34 kg-cm) ((weight = volume (3.52in3) * density
Figure 43: Hitec HS-5585MH
(0.723oz/in3), (torque = center of mass (~1.86in) *
weight (2.54oz))), resulting in a total required
torque of ~600 oz-in (~43.2 kg-cm).
Based on these calculations, the decision was made to use the Hitec HS-5585MH digital
servo (HS-5585MH Economical, High Voltage, High Torque, Coreless, Metal Gear Digital Sport
Servo, n.d.). Rated for 17 kg-cm of torque at 7.4V, they provide the needed torque, are quite
precise with a rated resolution between 0.079°/µsec to .134°/µsec, and are well reviewed. With a
configuration consisting of four motors with the pinky and ring fingers being actuated by the same
motor, the system creates 68 kg-cm of torque, allowing for any additional forces to be handled
while still achieving the desired load.
4.5 Control System
A prosthetic that requires any sort of input from its user must be intuitively and reliably
operable. Intuitive input from the user is currently being pursued in the industry using myoelectric
sensors near the location of amputation to pick up signals that would have once controlled the
motion of the, now amputated, body part. However, as described in the background, effective
myoelectric sensing technologies are currently only available in very expensive prosthetics, and
thus budget prosthetic designers usually resort to clunky mechanical and body motion inputs.
The control system developed for our prosthetic aimed to combine very simple myoelectric
and mechanical inputs with dynamic force sensing technology in order to provide the user with a
simple yet effective interface with their prosthetic. In short, the user need only select the type of
grip they desire and send a myoelectric signal to begin performing that grip. Force sensors located
on the fingers and palm of the hand take care of how much force the hand will apply to the object
being gripped and will dynamically adjust this force to avoid failure of slippage. The user then
need only send another myoelectric signal to tell the hand to open. This section outlines the design
and components of various blocks within our proposed control system. A view of the entire system
can be found in Appendix E.
47
4.5.1 DC to DC Converter
The power requirement for the servos used in this project called for 7.4V batteries.
However, the microcontroller used requires an input voltage of 5V, so it was necessary to design
a DC to DC converter to bring the voltage from the battery down to an acceptable level. The
LM1084 voltage regulator is a low dropout IC that has the ability to output a steady 5V with the
right voltage divider applied across its ground. The datasheet indicates that the output is
determined by the following equation:
[]
 = 1.25 (1 +
2
)
1
Using resistances of R2 = 330Ω and R1 = 1kΩ in equation (1) results in a VOUT of 5V. The
input to the converter is a female barrel jack and there are capacitors between Vcc and GND to
act as filters. Finally, a small LED is used to indicate that the device is on. The schematic below
shows the schematic for the complete converter.
Figure 44: Converter Schematic
4.5.2 Myoelectric Input Block
The control system calls for two easy to produce signals from the user to denote a desire
to close the prosthetic hand and then to reopen it. It is not necessary for the system to interpret
what grip the user desires as that will be handled by a mechanical input described further in the
report. The MyoWare Muscle Sensor is a prefabricated PCB that contains the necessary
hardware to translate small variations in voltage in a muscle into analog values that can then be
read by a microcontroller. Our control system uses two Myoware sensors placed on the flexor
muscles of the upper forearm.
48
The MyoWare Muscle Sensor receives its signals from electrodes that are in direct contact
with the user’s skin. There are many different types of electrodes, but this design will focus on
multi-purpose medical electrodes and conductive fabric. It is best
that the electrodes receive a strong, clear signal. This can be
achieved by maintaining skin-to-surface contact, receiving minimal
noise and interference, and having a high conductivity. Prototyping
of the muscle sensor consisted solely of the use of electrodes as
they provide a much clearer signal compared to conductive fabric.
The schematic shows a 6-pin configuration for the
myoelectric sensors. Two pins are used as power, two are used as
ground, and A1 and A2 are the respective analog output for each
sensor.
Figure 45: 6-pin configuration
4.5.3 Grip Selection
The second component that assists in user input is the grip selection block of the control
system. Using a rotary encoder, the user can quickly scroll to the desired grip they wish to perform.
When idling, the display will show the current grip that will be performed if a signal is received
from the user. Turning the encoder causes the display to show all the available grips mapped as
numbers between 1 and 5. The user need only push down on the dial after scrolling to their
desired grip and the display will hold that grip for execution.
The grips were mapped as thus in accordance with the possible finger positions described
in Section 4.1.
Table 5: Finger Positions
Grip #
Grip Name
Grip Description
1
Rest
For use when hand is not in use, fingers slightly curled
2
Point
All fingers closed except index (thumb closes last)
3
Rod Grip
Four fingers closed, thumb does not move
4
Pinch Grip
Index, middle, and thumb finger meet
5
Spherical Grip
All five fingers close together
Figure 46 below shows the schematic for the grip selection block. The rotary encoder is
modeled as a potentiometer that feeds an analog signal into the microcontroller. The display is
modeled as a common anode display driven by a BCD to seven-segment decoder. The decoder
minimizes the amount of pins necessary to drive the pin-heavy display by decoding a four digit
binary signal. Resistors are used to limit the current flowing into the display so as to not burn out
the LED’s inside the display.
49
Figure 46: Grip Selection Block Schematic
Figure 47: Force Sensor Array Schematic
50
4.5.4 Actuation Feedback Block
It was necessary to design a method for gathering
force data on the hand for the microcontroller to make
decisions on actuation. Without such feedback, the motors
driving the fingers would simply actuate as much as
possible and stall when the fingers cannot move any
longer. This does not allow for fine motor control, which
would not be beneficial if the user wanted to, for example,
shake someone's hand.
The actuation feedback block of the control system
uses a series of force sensitive resistors that are be placed
all over front of the palm and fingers. These resistors, when
squeezed, allow additional voltage through them, and this
difference in voltage can be measured by a
microcontroller. In order to interface with the necessary
Figure 48: Sensor Placement
amount of force sensors to cover the whole hand, shift
registers were utilized to cycle through and read each resistor one at a time. This process of
reading each resistor takes only about 100ms (or 1/10th of a second). The advantage of this
circuit is that it only requires a few digital pins to run the shift registers and one analog input for
all of the resistors.
The schematic below shows the final construction for the force sensor array. The three 6pin connectors labelled J4, J5, and J6 are where the force sensors will be connected.
The microcontroller can use the force sensor to determine if the servos need to be turned
more or if they can be locked in place. As long as the user does not send an open-hand signal,
the force sensors will stay active, checking if more force is needed from the servos to maintain a
firm grip. Twelve resistors are used to effectively cover the area of the hand and are arranged as
pictured in Figure 48.
4.5.5 Servo Control Block
Due to the decision to use servos as opposed to other forms of actuation to actuate the
hand, this block of the control system is very simple. Servos have a built-in microcontrollers as
well as position tracking that can be interfaced with using only a serial data signal. This means
that the setup of the servo control block simply consists of supplying power to the servos at their
rated voltage and connecting the servo digital pins to the microcontroller. The servos can be
moved using a system called pulse width modulation (PWM). PWM refers to the concept where
short, measured bursts of voltage are used to encode specific control signals. For example, these
servos will move to their center position when a 1.5ms pulse is sent. The schematic for this block
can be seen below.
51
Figure 49: Servo Control Block
The servo’s chosen for this block require 7.4V, which became the constraining voltage for
the entire system. It is also important to note that the ground for this block, as well as all other
blocks, is shared with the microcontroller. This allows both processors in the servo and the
microcontroller to see the same signal when messages are passed between them.
4.5.6 Microcontroller
One all of the components of the control system were determined, the decision for a
microcontroller was made based on power requirement and number of inputs and outputs
available. By using decoding and multiplexing techniques in both the actuation feedback and grip
selection blocks, the overall amount of pins necessary to handle all of the blocks was minimal.
Thus, the ATMega328P was chosen due to its ability to use a 16MHz clock and its relatively low
power requirement.
A large portion of control is being performed by the microcontroller in the form of constantly
looping code that takes input from the three input blocks and outputs the corresponding signals
to the servo control block. The flow diagram for this code can be seen below. It effectively consists
of a phase of three states: a calibration state, an idle-open state, and an idle-closed state. The
calibration state is described in the section below. The idle-open state waits for a myoelectric
signal to start closing the fingers, and begins actuating the fingers when the signal is received.
The code then moves into the idle-closed state where it continuously determines whether there
is a solid grip being applied to an object or if maximum actuation has been reached. While it is
checking the force sensors, it also waits for a signal to move back into the idle-open state. When
this signal is received, the motor are instructed to actuate in the opposite direction, opening the
hand to its full extension.
52
Figure 50: Microcontroller
The code as a whole operates on an
interrupt-driven 100Hz clock. Each time an interrupt
occurs, the interrupt service routine (ISR) sets
Booleans that act as triggers for the sensors to be
polled and an appropriate action based on the state
machine described above in the main loop to be
taken, individually. Polling the sensors involves
reading and storing the values read from each force
sensor and the Myoelectric sensor. These are then
used for determining the active state’s action. For
the idle-open state, the Myoelectric sensor is used
to determine if the hand should start closing. Also,
during this state, the rotary encoder is being polled
to determine which grip will be selected once the
Myoelectric sensor value threshold has been
exceeded. During the closing state, the force sensor
readings are used to determine if an object is either
blocking or in the grasp of a finger and that finger
should stop moving. This is accomplished by writing
each servo a maximum of five degrees further
towards its desired state from the previous state until
the finger’s respective force sensor reaches its
threshold reading or the finger reaches its desired
Figure 51: Logic Flow Chart
53
position. Once one of these conditions are met, a Boolean is set stating that the finger is finished
moving. This is both used in the event of the thumb/fingers moving asynchronously due to the
requirements of the grip as well as to move to state 3, idle-closed. The idle-closed state acts very
similarly to the closing state in that the servos are still told to move in order to adjust for slip from
grasped objects. The primary difference is that another “high” reading from the Myoelectric sensor
will cause the fingers to open and the state to change to idle-open once the fingers are open.
4.5.7 Battery Selection
In order to drive all the electronics, an appropriate power source must be selected. It needs
to have enough voltage to drive the servos as well as enough charge (measured in mAh) to power
all the electronics for an extended period of time. In particular, the battery needs to be rated for
7.4V and have enough charge to last through a day of standard usage. In order to determine this,
amperage values for different states of the finger servo were measured and labelled below as Idle
(no motion), Stall (servo is attempting to move but cannot), and Average (servo is moving with
varying force applied). The amount of time the servo could operate within these three states was
calculated for two different battery. The amount of time four servos could operate was also
calculated to mimic the actual use case within our system.
Table 6: Estimated Battery Life Based on Servo Power Draw
Current Draw
Battery
Amount of
Motors
5000mA-hr
LiPo Battery
4600mA-hr
LiPo Battery
Idle ~= 5mA
Stall ~=2500mA
Average ~= 200 - 1000 mA
1
1000 hours
2 hours
25 hours - 5 hours
4
250 hours
0.5 hours
6.25 hours - 1.25 hours
1
920 hours
1.84 hours
23 hours - 4.6 hours
4
230 hours
0.46 hours
5.75 hours - 1.15 hours
Appendix C.2 indicates that on average a person will use their hand to perform basic ADL’s
for about 5 hours and 6 minutes every day. The following calculations determine two different
mA*hr ratings using Idle and Average amperages:
[] ℎ 
= (   ∗   )
+ (     ∗   )
[] ℎ  () = (19ℎ ∗ 5) + (5ℎ + 200) = 1095ℎ
54
[]
ℎ  () = (19ℎ ∗ 5) + (5ℎ + 1000) = 5095ℎ
If the maximum average current is used, the amount of mA*hr necessary is just above the
rating of the first battery, and this isn’t accounting for when the prosthetic is off at nighttime.
However, it became necessary to choose a smaller battery due to size constraints of our
prototype. As such, the lower mA-hr battery was chosen for our prototype.
4.5.8 Myoelectric Data Processing
The ultimate goal of the data processing steps applied in this project is to accurately
predict the user’s desired action and act accordingly. A Kalman Filter is used to filter the incoming
data which is then sent to an expectation maximization algorithm, a form of data clustering
algorithm, to accurately determine the user's input. After being sent to the EM algorithm, the data
is clustered based on its sensor value and change in sensor value to determine if the user has
the corresponding muscle relaxed, flexed, or if he/she is in the process of flexing or relaxing. This
is then sent to the state machine to generate an appropriate output. As an example, if the
algorithm determines the user is starting to flex with the hand in an open position, it would start to
close the hand.
The first stage of the data processing cycle is a Kalman Filter. The signal from Myoelectric
sensors is inherently noisy. This, in turn, can make predicting the user’s actions more difficult at
a single point in time or requires sampling and averaging over 100-200 milliseconds (ms) period
of time. A way to circumvent this problem is through filtering the stream of data. As such, a
position-velocity (PV) model Kalman Filter was implemented as a software filter for this purpose.
“A Kalman filter is an optimal estimator - i.e. infers parameters of interest from indirect, inaccurate
and uncertain observations…. [It is optimal in the sense that if] all noise is Gaussian, the Kalman
filter minimizes the mean square error of the estimated parameters” (Kleeman, n.d.). As seen in
Figure 52, the Kalman Filter smooths the sensor data quite well while representing that data
relatively well.
Figure 52: Kalman Filter on Muscle Sensor Data
55
Once run through the Kalman Filter, this data is then analyzed to create a mixture model,
a mathematical (probabilistic) model of distributions within a given dataset or input. From this
point, the data is then fed into an Expectation Maximization (EM) algorithm based on the mixture
model in order to cluster the data into categories of predicted actions. Particularly, the algorithm
utilizes two clusters: a Gaussian-based one representing the relaxed state and an exponentialbased one representing the flexed state. This is done in order to allow for greater prediction
capabilities than merely setting arbitrary thresholds for given actions. Furthermore, it creates an
effective way of calibrating the hand to the user when combined with a calibration process when
the hand is first used. The results of an early EM prototype can be seen in Figure 53.
Figure 53: Results of Early EM Algorithm Prototype on Filtered Sensor Data
56
5. Testing and Results
5.1 Tests
In order to assess the functionality of the team’s prosthesis, several tests were developed
and carried out. Each test had a specific parameter it was assessing, and the standards and goals
were drawn from the product specifications. The prosthesis was tested as a system, its individual
components were tested when necessary. For full test procedures, see Appendix F: Full Test
Procedures.
5.1.1 Pin Fracture
The objective of this test is to determine whether the pins placed at the finger joints were
able to allow full motion of finger and hold the required load without failing. The load the pins in
the joints and hand will have to withstand a maximum load of 11lb +/- 1lb. The goal of the
prosthetic hand is to allow a user to be able to complete most activities of daily living. Through
research it was found that most activities can be completed by being able to lift 11 pounds. If the
pins are able to hold up to 11 pounds then that would mean the pins are successful and can be
used within the prosthetic.
5.1.2 Thread Ductility
The objective of the test is to determine at what weight the Kevlar thread would deform.
The Kevlar threads are what allow the hand to actuate. They need to be able to withstand the
forces that will be placed on the fingers and hand. All the fingers together will pick up objects,
but to ensure the hand will be able to hold any object the fingers will be tested individually. The
Kevlar thread needs to be able to support at least 11 pounds. The test is also done to determine
how long before the Kevlar will deform and elongate. When the Kevlar thread elongates it
makes it more difficult to pull on the thread with the servo. It is important to understand when the
thread will deform and when it needs to be serviced.
5.1.3 Hand Friction
The 3D printed hand must be capable of withstanding the wear and resisting the friction
the user subjects it to. The hand includes all pins, finger components, palm assembly, and the
O-rings that aid actuation. Understanding how the hand will respond under repeated and
continuous movement will determine if the hand is ready to be used in the final assembly as it is
designed. Failure to withstand the wear means a redesign will be necessary.
The objective of this test is to determine whether the hand will be able to withstand
continuous and full actuation. The test will be ran for 600 cycles for a fingers. One cycle is
defined as the finger actuating from its open position to its closed position and back to the open
position. The classification of success and failure is described below.
57
5.1.4 Socket Slippage
The objective of this test is to measure the slippage of the prosthesis on the arm. To
simulate normal loading and unloading during daily use, the prosthesis will be affixed to a model
of a residual limb as if it were an actual amputee’s arm and tightened securely. Hanging masses
will be hung from the prosthesis for one minute. Multiple angles will be tested. The starting and
ending positions of the prosthesis will be marked and measured. A Likert Scale will be used to
rank the level of success or failure after the five minutes have passed.
To test the slippage of the socket, the socket will be attached to the model of the residual
limb at one of three test angles: 0 degrees (horizontal), 45 degrees, and 90 degrees (vertical).
Hanging masses in the form of pre-weighed volumes of water are to be hung from the fingers.
Marked slip distance after one minute of testing will be measured with calipers that have been
calibrated properly.
To test slippage on human skin, the socket portion of the prosthesis with the base plate
removed can be worn on an actual human arm. The same test masses can be hung and the test
can be repeated.
5.1.5 Finger Actuation
The objective of this test is to determine if the design of the finger was able to complete
the full motion required to complete desired activities. The test will see if using the developed
program the fingers will be able to open and close. The amount of friction will be tested between
the joints to ensure there is enough room for each joint to open and close. The servo will also be
tested to see how many cycles it can withstand before any complications arise. The different
components will also be tested to see when they need to be serviced. Where the thread is
attached to the servo can deteriorate over time and will be observed to see how many cycles it
can withstand before breaking.
5.1.6 Control System Hardware
The objective of this test is to verify the functionality and power specification of the
electrical system that actuates the prosthetic. All blocks within the system must be able to function
simultaneously from the same power supply and under maximum current draw. The test will
explore every possible usage of the system and determine whether it is acting in accordance with
its proposed function. The tests within this section are separated into software and hardware tests
in order to isolate errors.
5.1.7 Control System Software
For the prosthetic to function properly, each individual system needs to function in of itself
as well as a whole. Since the prosthetic is controlled by software, this software needs to be tested
to ensure this functionality. The code written must be robust, reliable, and relatively bug-free while
interacting with the other physical systems within the prosthetic effectively.
The objective of this test is to determine if the written software and control systems can
appropriately operate each part of the prosthesis. The software must be able to interact with the
mechanical and electrical systems. These tests explain various tests to ensure the integrity of
these interactions.
58
5.1.8 Hand Functionality
The objective of this test is to assess the functionality of the prosthesis’ pre-programmed
grips and gestures. These are: Open Hand, Closed Hand, Point, and Pinch. To do this, a scale
of 1-7 is used to determine how well the prosthesis accomplished an assigned task, with 7 being
success and 1 being failure.
For this test, the prosthesis should be either secured to a test subject’s arm (note: the
base plate should be removed if the subject does not have a residual limb), or to a model of a
residual limb. Once secure, the prosthesis will be tested using the procedures and objects
outlined below. Each grip/gesture is to be tested with different objects as appropriate.
5.2 Test Results
Below is the results of the test procedures carried out by the team. Each section notes a
different test procedure, the results, and the notes taken throughout the test session.
5.2.1 Pin Fracture
The Pin Fracture test, Appendix F.1, was designed to ensure that the pins used in the
joints could withstand the load of weight up to 10 lb. The test was run by placing a finger segment
fully assembled, with a pin and washers to recreate the application as best as possible, in an
Instron machine. The assembly was tensioned until failure. The test was marked a success as
the pins tested far exceeded the 10 lb. requirement. The results can be found in Appendix F.1.
5.2.2 Thread Ductility
The Thread Ductility test Appendix F.3, was designed to ensure that the Kevlar thread
would not elongate and plastically deform to a length that would cause the finger to be unable to
actuate completely with the strength it was designed to perform. The test was run by hanging 10
lb. from the end of three lengths of thread and measuring the original and then the elongation.
The test was marked a success as the thread did not elongate more than 3%. The results can be
found in Appendix F.3.
5.2.3 Hand Friction
The Hand Friction test was designed to ensure that the hand used in the prosthetic did
not wear down to inoperable dimensions. The test was run by measuring each component used
in a finger, actuating the finger through 500 cycles and remeasuring the dimensions. The
components could not wear down more than .05 inches from its original measurement. The test
was marked a success as all components fell into the range they are deemed as functioning.
5.2.4 Socket Slippage
The Socket Slippage test, Appendix F.4, was designed to ensure that the socket used in
the prosthetic was functioning as designed. The test was run by monitoring how far the socket
59
would slip down the stump when exposed to weight at varying angles. Readings were taken with
a pair of calipers as weight was increased up to 10 lb. The test was marked a success as the
socket did not slip past the limit of 15 cm. The results can be found in Appendix F.4.
5.2.5 Finger Actuation
The Finger Actuation test, Appendix F.2, was designed to ensure that the finger designed
in the prosthetic actuated properly. The test was run by actuating the finger using the designated
servos. The angle was measured of each component to ensure the finger actuates evenly and
completely. Additionally, the finger was required to actuate within 2 seconds to be deemed
responsive and capable of normal grip function. The test was marked a failure as the fingers were
not actuating properly. The fingers have worked in the past but an assembly error was made prior
to the test with the threads that caused the failure. The test was re-run and passed the test. As
the failure was due to an assembly error, the team has deemed the test an overall failure as
repeatability of the passing run may not easily be repeated by users.
5.2.6 Control System Hardware
The Control System Hardware test, Appendix F.5, was designed to ensure that the
hardware used in the prosthetic was functioning as designed. The test was run by monitoring
each port of connection in the hardware. Readings such as voltage, current, and signaling were
taken. The test was marked a success as all connections and readings fell into the range in which
they are deemed as functioning. The results can be found in Appendix F.5.
5.2.7 Control System Software
The Control System Software test was designed to ensure that the software code used to
control the prosthetic as designed. The test was run by monitoring each function of the code so
that it communicates and controls the appropriate component of the prosthetic properly.
Observations of the input, output of both the electrical and physical hardware were taken to
determine if they were successful or have failed.
5.2.8 Hand Functionality
The Hand Functionality test, Appendix F.6, was designed to ensure that the hand used in
the prosthetic could complete the actions it was designed to do. The test was run by actuating the
hand through different grips patterns, holding daily-use objects such as doorknobs and water
bottles, etc. The hand must complete all these functions and tasks to be considered a success.
The test is currently marked incomplete, which deems the test a failure. This is due to the fact
that finger actuation was ultimately deemed a failure due to repeatability concerns. The team
believes with the right preparation that the test would succeed but did not find it possible for users
to repeat the actuation success that was necessary to pass the functionality test.
60
6. Project Expansion and Continuation
Over the course of the project, this MQP Team has designed a functional socket, forearm
servo housing, hand, and fingers from the ground up. Best practices from previous projects were
adapted where necessary. As a system, the goal up to now was to prove that the team’s designed
system can operate appropriately, opening and closing the fingers at the user’s input. The
prosthesis will open its fingers, and then close around an object until the force sensors trigger,
causing the fingers to stop.
However, throughout this entire process many other features to implement for the
prosthesis were considered and in some cases thoroughly explored. Most of these features seek
to create a prosthesis that more closely mimics the functionality of a real hand. Prototype
implementation and testing of these features was not carried out due to time and specification
constraints.
6.1 Haptic User Feedback
“For a myoelectric prosthetic user, efferent signals [haptic feedback, in other words] on
the residual limb are used to control the prosthesis; however, the prosthetic device does not
compensate for the loss of afferent signals. This requires the amputee to rely on vision alone for
precise control of the prosthesis reducing the effectiveness and speed at which it can be operated.
… The importance of incorporating a sensory substitution device in a prosthesis is becoming
increasingly essential as more functional hand prostheses are being developed with multiple
degrees of freedom (DOF) as compared to 1 DOF basic gripper prostheses” (Erwin & Sup, 2015).
As such, a concept in development for this project is using air-filled sacks that vary in pressure
depending on the amount of force detected by force sensors in several locations on the hand.
Depending on which sensors detect changes in force, corresponding sacks will inflate or deflate
in response. The air sacks themselves would be placed along the arm similar to a blood pressure
cuff. Small fish tank air pumps would be used to inflate these air sacks along with solenoid driven
release valves to reduce pressure once the force sensors no longer detect force along the hand.
Overall, this system is meant to give the user a sense of feeling based on what the prosthetic
hand is experiencing.
6.2 Cooling
A 2001 study of nearly 100 amputees found that heat and perspiration inside the socket
were reported by 72% of the survey participants as the most common cause for a reduced quality
of life. Removing perspiration from the surface of the skin and developing an active cooling system
that is capable of reducing residual limb temperature to decrease perspiration and increase socket
comfort72, are necessary for prosthetic sockets.
Four alternative cooling system designs were identified: fluid, computer central processing
unit (CPU) fans, CPU fans combined with a heat sinks and peltiers, and open frame socket. These
72
Farrell, 2016
61
designs were evaluated against the ranked design goals in Table 7 below. It was decided that the
best cooling system is CPU fans followed closely by CPU fans with heat sinks and peltiers. Open
frame sockets aren’t really a method of cooling but that they are cooler than typical sockets
because they don’t retain as much heat. Therefore, all three of these methods will be tested and
compared.
Table 7: Cooling System Decision Matrix
Cooling System Decision Matrix
Design Alternatives
Design Goals
Weighing
Factors
Fluid
CPU
Fan
CPU Fan, Heatsink, &
Peltier
Open
Socket
Safety
16
7.5
8.5
6.5
9
Ease of Use
15
8.5
10
10
8
Durability
14
7
10
10
10
Comfort
13
8.5
5
8
6
Range of Motion
11
7.5
9
9
9
Max Capable Load
10
NA
NA
NA
NA
Service-ability
9
3
8
7
8
Weight
8
1.5
9
8
7
Scalability
7
7
10
10
10
Cost
6
3
8
5
5
Difficulty to
Assemble
5
3
9
8
5
Enviro. Factors
4
10
5.5
7.5
8
Aesthetics
3
NA
5
5
7
Time to Assemble
2
3
9
7.75
5.5
Time to Print
1
NA
NA
NA
5
Total
674
918.5
890
870
Note: The decision is based upon unknowns and the design alternatives are not comparable in every category.
An option to supplement or replace the Open Socket design is through the use of computer
central processing unit (CPU) fans. This design is in the comparison phase. CPU fans would act
as an active form of cooling to further cool areas of contact that the Open Socket would otherwise
62
not cool as well due to its passive nature. One or more CPU fans, depending on the size of the
fans and space constraints of the afflicted area, would be placed along areas deemed to be “high
heat” areas. Certain areas result in more heat and friction and could use supplemental cooling.
CPU fans could, in themselves, also be supplemented through the use of peltiers. Peltiers
operate on the principles of thermoelectric cooling. “All electric current is accompanied by heat
current (Joule heating). What [Jean] Peltier [discoverer of the Peltier Effect] observed was that
when electric current passed across the junction of two dissimilar conductors (a “thermocouple”)
there was a heating effect that could not be explained by Joule heating [heat being generated
from electricity passing through conductors] alone. In fact, depending on the direction of the
current, the overall effect could be either heating or cooling. This effect can be harnessed to
transfer heat, creating a heater or a cooler”.73 This concept creates a temperature differential
across the device. One side gets hot while the other gets cold. Due to this, peltiers could be used
in conjunction with CPU fans and heatsinks in order to cool the user’s limb and remove the heat
from the peltier. This design is in the comparison phase.
The concept of thermoelectric cooling has also been used in a design by Leto Solutions
for a prosthetic leg.74 Though likely different in execution from what this project’s arm would likely
require, this shows the viability of the use of the technology. However, how well the Open Socket
reduces the amount of heat experienced by an amputee will need to be further explored to
determine the necessity of either CPU fans or peltiers.
6.3 Wrist
In our design the wrist would not be in the location of a typical human arm, rather it would
be fairly close to the site of a transradial amputation. The location of actuation has been
intentionally relocated to mitigate the complications that would result by having stationary servo
motors operate rotating fingers.
A system to have the arm be able to rotate 180 degrees will be created and connected to
the forearm servo housing and the wrist servo housing. The system that would be implemented
is a planetary gear system and a servo. The planetary gear system will use spur gears that will
have a ratio of 3:1 from ring to sun gear and will have an external diameter of two inches. The
outer diameter of the largest component, the socket frame, is currently three inches. Therefore,
the system that will be used to turn the arm should also be roughly the largest diameter there
currently is. It must be able to rotate the prosthetic while at the same time meeting its size
requirement.
When selecting a servo for wrist rotation, as with the finger motors, the goal is to have the
capability of moving around 8-10 pounds, or approximately the weight of a gallon of milk to meet
the requirements we set. As such, the dimensions of a milk jug in order to accurately determine
how where the center of gravity is of a milk jug and how much torque would be exerted on the
arm from it, determined to be a minimum of 400 oz-in of torque. This would require the use of a
gearbox to amplify the torque of the selected motor. A drawback of using this in conjunction with
73
74
Understanding Thermoelectric Cooling, n.d.
Pace, 2014
63
a servo, though, is the fact that a servo’s range of motion is generally limited to between 180-360
degrees, depending on the servo. Using the gearbox would further reduce this range of motion
and would likely result in not being able to achieve a full 180-270 degree rotation to mimic a
human wrist. To circumvent this problem, a continuous rotation a possible servo was selected to
be used.75 A continuous rotation servo was chosen due to the fact that the gearbox would limit
the maximum amount of rotation achievable by a normal servo. This motor will then be mated in
a 3:1 ratio gearbox to increase its torque to the needed amount.
Though by moving the wrist up the forearm the group has removed many issues, one still
exists - the muscle contraction sensing wires will have to remain stationary on the user’s muscle,
while rotating along with the prosthetic as the wrist is turned. A solution the group is considering
for this problem is a conductive track which runs around the circumference of the prosthetic to
allow for a stationary muscle contraction wire to be in constant contact with the moving arm.
Figure 54: Possible design for wrist
6.4 Scalability
Equations and Relations is a powerful tool SolidWorks provides to its users. It is a duty of
the team to harness the power of this feature and make use of its functionality to simplify the
models design process and give the hand the ability to be scaled to different sizes for different
users. Based upon the 2015-2016 3-D Prosthetic Hand MQP, a few relations have been identified
in the palm and fingers which will used in this project. The group had great success in defining
and creating these relations. It could be an element that adds to the unique qualities of the group’s
product.
75
HSR-2645CR Servo, n.d.
64
Figure 55: Equations and relations table
The equations and relations have been broken into their locations, the palm and the threejoint fingers. The palm contains relations such as lengths and heights that define separate
dimensions of the palm which are all dictated by the “inputhandwidth” relations which has a
domino-effect on the other relations. The fingers contain similar, yet more uniquely attributed,
equations all dictated by a similar “inputhandwidth” command that exists by relating to the palm’s
relations. These relations are all reliant, as well as commanded by one another with an order of
command that trickles through the data flow. With these being identified, the hand itself will have
the powerful option of scaling for different users in terms of their personal physical measurement
fittings. This way, users of all ages and varying physical attributes the opportunity to adopt the
product.
Even with the relations from last year, the team must define their own relations to their
model. Last year’s model was not altered to create the new product due to the difference in nature
and goals of the projects. Beginning with a new model allows for the ability to adopt overarching
relationships from last year and gives the group the opportunity to define their own relations.
6.5 Engineering Standards
With the development of a new product there has to be standards in place to make sure
that it can be compared to other products and see if it can even be used commercially. The
standards that are used currently are in the testing aspect of a prosthetic. The prosthetic must
have all of its components undergo stress tests to see if they can support enough weight to serve
as a prosthetic. In addition, the device must undergo cyclic testing to see how long it can function
without breaking. For prosthetics in general the cyclic testing is very important and while there is
no standard for prosthetic hands of this caliber currently using similar parameters would be
acceptable. If further development of the prosthesis is done then recording the data from the cyclic
and load tests in compliance with the standards listed below.
65
E467 - 08(2014)
E1942 - 98(2010)e1
E2443 - 05(2010)e1
E647 - 15e1
E1457 - 15
E1681 - 03(2013)
E2760 - 16
Standard Practice for Verification of Constant
Amplitude Dynamic Forces in an Axial Fatigue
Testing System
Standard Guide for Evaluating Data
Acquisition Systems Used in Cyclic Fatigue
and Fracture Mechanics Testing
Standard Guide for Verifying ComputerGenerated Test Results Through The Use Of
Standard Data Sets
Standard Test Method for Measurement of
Fatigue Crack Growth Rates
Standard Test Method for Measurement of
Creep Crack Growth Times in Metals
Standard Test Method for Determining
Threshold Stress Intensity Factor for
Environment-Assisted Cracking of Metallic
Materials
Standard Test Method for Creep-Fatigue
Crack Growth Testing
References
3M. Multi-purpose monitoring electrode. Retrieved from https://www.amazon.com/Red-DotMulti-Purpose-MonitoringElectrode/dp/B002C2M90I/ref=pd_sim_200_1?ie=UTF8&pd_rd_i=B002C2M90I&pd_rd_
r=AHTRQTN860H5F0SCBQXB&pd_rd_w=ANTxp&pd_rd_wg=iCweR&psc=1&refRID=A
HTRQTN860H5F0SCBQXB
66
Adafruit. Knit jersey conductive fabric. Retrieved from
https://www.adafruit.com/products/1364?gclid=CjwKEAjw34i_BRDH9fbylbDJw1gSJAAvI
FqUlxMTFSjmapWFwG5yhf9lup6E0-w7fQ0GOSyHuldzQBoCPxPw_wcB
Ada Hand Kit. (n.d.). Retrieved September 7, 2016, from http://www.openbionics.com/shop/ada
Advancer Technologies. MyoWare muscle sensor. Retrieved from
http://www.advancertechnologies.com/p/myoware.html
Alley, L. R., Williams, T. W., Albuquerque, M. J., & Altobelli, D. E. (2011). Prosthetic sockets
stabilized by alternating areas of tissue compression and release. Journal of
Rehabilitation & Development, 48, 679-696.
Amputee Treatment Center. Pump-It-Up! socket system. Retrieved from http://www.amputeecenter.com/pump.htm
Arduino - ArduinoBoardMega2560. (n.d.). Retrieved October 09, 2016, from
https://www.arduino.cc/en/Main/ArduinoBoardMega2560
Bailey, Aubrey. "What Are the Different Types of Prosthetics?" What Are the Different Types of
Prosthetics? Chron.com. Chron, 2012. Web. 13 Dec. 2016.
Behrens, J. (n. d.) Structure and function of the hand. Mercer County Community College
Physical Therapy Assistant Program. Retrieved from
http://www.mccc.edu/~behrensb/documents/Structurefunctionofthehandbjb.pdf.
Beil, T., Street, G., and Convey, S. (2002). Interface pressures during ambulation using suction
and vacuum-assisted prosthetic sockets. Journal of Rehabilitation Research and
Development, 39, 693-700.
Bio Designs. (n.d.) HiFi Interface System.
http://www.biodesigns.com/interface_socket_design.html
BNC Connector. (n.d.). In PC Magazine Digital Edition Encyclopedia online. Retrieved from
http://www.pcmag.com/encyclopedia/term/38800/bnc-connector
Brown University. (2003). Hand Prosthetics: Statistics. Retrieved from
http://biomed.brown.edu/Courses/BI108/BI108_2003_Groups/Hand_Prosthetics/stats.ht
ml
Click Medical. (n.d.). Revofit. Retrieved from https://www.clickmedical.co/store/revofit/revo/
Commean, P. K., Smith, K. E., & Vannier, M. W. (1997). Lower extremity residual limb slippage
within the prosthesis. Archives of Physical Medicine and Rehabilitation, 78, 476-485.
Daly, W., Voo, L., Rosenbaum-Chou, T., Arabian, A., & Boone, D. (2014). Socket Pressure and
Discomfort in Upper-Limb Prostheses: A Preliminary Study. Journal of Prosthetics and
Orthotics, 26(2), p. 99-106.
Delph, M. A., II, Fischer, S. A., Gauthier, P. W., & Martinez Luna, C. H. (2012). Rehabilitative
Robotic Glove (Rep.). Retrieved September 25, 2016, from WPI Automation and
Interventional Medicine Laboratory website: https://web.wpi.edu/Pubs/E67
project/Available/E-project-042512143542/unrestricted/Rehabilitative_Robotic_Glove_MQP.pdf
Department of Orthopedic Surgery, San Francisco General Hospital (2011). Transradial
Amputation. Retrieved from http://orthosurg.ucsf.edu/oti/patient-care/divisions/orthoticsand-prosthetics/transradial-amputation/
Devadass, V., Adnan, J., Mohamed, N. A., Razikin, M. F., Aziz, A., & Hisham, M. N. (n.d.). A
step forward in below knee prosthesis socket design. Materialise. Retrieved from
http://biomedical.materialise.com/cases/step-forward-below-knee-prosthesis-socketdesign
DivideWorks. (n.d.). Making InMoov Move. Retrieved September 10, 2016, from
http://www.instructables.com/id/Making-InMoov-Move/?ALLSTEPS
Dorenfeld, Elyssa , Wolf, Robert , Zeveska, Stephan (2013). Design of a Powered Hand
Orthosis.
Retrieved from: http://digitalcommons.wpi.edu/atrc-projects/22
Erwin, A., & Sup, F. C. (2015). A Haptic Feedback Scheme to Accurately Position a Virtual Wrist
Prosthesis Using a Three-Node Tactor Array. PLOS ONE PLoS ONE, 10(8).
doi:10.1371/journal.pone.0134095
Farrell, T. (2016). Development of Moisture Management Liner and Active Cooling System for
Improving Residual Limb Skin Care. Liberating Technologies: Research. Retrieved from
http://www.liberatingtech.com/research/#20131019_154757
Ghoseiri, K. & Safari, M. R. (2014). Prevalence of heat and perspiration discomfort inside
prostheses: Literature review. Journal of Rehabilitation Research and Development,
51(6), 855-868.
Greenwald, R. M. (2003). Volume Management: Smart Variable Geometry Socket (SVGS)
Technology for Lower-Limb Prostheses. Journal of Prosthetics and Orthotics, 15(3), 107112.
Hedef (n.d.). Retrieved from http://www.hedefprotez.com/tr/urunler/item/oessur-seal-in-x5.html.
Highsmith, J. T., & Highsmith, M. J. (2007). Common skin pathology in LE prosthesis users.
JAAPA: Journal of the American Academy of Physician Assistants, 20, 33-36, & 47.
Hitec 35485S HS-5485HB Standard Karbonite Gear Digital Top BB Servo. (n.d.). Retrieved
October 1, 2016, from https://www.amazon.com/Hitec-HS-5485HB-Standard-KarboniteDigital/dp/B001KYSE2G
Hitec 37954S HS-7954SH High Voltage Ultra Torque Dual BB Servo. (n.d.). Retrieved October
01, 2016, from https://www.amazon.com/Hitec-37954S-HS-7954SH-VoltageTorque/dp/B0027L7UDA
HS-5585MH Economical, High Voltage, High Torque, Coreless, Metal Gear Digital Sport Servo.
(n.d.). Retrieved October 01, 2016, from http://hitecrcd.com/products/servos/sport68
servos/digital-sport-servos/hs-5585mh-economical-high-torque-digital-corelessservo/product
HS-5645MG High Torque, Metal Gear Digital Sport Servo. (n.d.). Retrieved October 1, 2016,
from http://hitecrcd.com/products/servos/sport-servos/digital-sport-servos/hs-5645mghigh-torque-metal-gear-servo/product
HSR-2645CRH Servo. (n.d.). Retrieved October 31, 2016, from https://www.servocity.com/hsr2645crh-servo
Kelso, T. (2015). More insight into developing grip strength: Your hand digits. Breaking Muscle.
Retrieved from https://breakingmuscle.com/strength-conditioning/more-insight-intodeveloping-grip-strength-your-hand-digits
Kleeman, L. (n.d.). Understanding and Applying Kalman Filtering. Retrieved December 10,
2016, from
http://biorobotics.ri.cmu.edu/papers/sbp_papers/integrated3/kleeman_kalman_basics.pd
f
Lake, C. (2008). The Evolution of Upper Limb Prosthetic Socket Design. Journal of prosthetics
and orthotics (1040-8800), 20(3), 85-92.
Lake, C., & Dodson, R. (2006). Progressive upper limb prosthetics. Physical medicine and
rehabilitation clinics of North America, 17(1), 49-72.
Legro, M. W., Reiber, G., Aguila, M., Ajax, M. J., Boone, D. A., Larsen, J. A., … Sangeorzan, B.
(1999). Issues of importance reported by persons with lower limb amputations and
prostheses. Journal of Rehabilitation Research and Development, 36, 155-163.
Mak, A. F., Zhang, M., & Boone, D. A. (2001). State-of-the-art research in lower-limb prosthetic
biomechanics- socket interface: A review. Journal of Rehabilitation Research and
Development, 38, 161-174.
Maguire, M. T., & Boldt, J. (2013). Prosthetic Rehabilitation Manual: Transfemoral (Above Knee)
Amputation.
Marchessault, J. A., McKay, P. L., & Hammert, W. C. (2011). Management of Upper Limb
Amputations. The Journal of Hand Surgery. 36(10), 1718-1726.
Martin Bionics. (n.d.). Socket-less socket: Fitting and fabrication training. Retrieved from
http://www.martinbionics.com/
Mathiowetz, V., Kashman, N., Volland, G., Weber, K., Dowe, M., Rogers, S., Arch Phys Med
Rehabil 66 (1985) 69-72
MSP432P401R LaunchPad. (n.d.). Retrieved September 24, 2016, from
http://www.ti.com/tool/MSP-EXP432P401R#tabs
McGimpsey, G., & Bradford, T. C. (2008). Limb prosthetics services and devices.
Bioengineering Institute Center for Neuroprosthetics, Worcester Polytechnic Institution.
National Commission on Orthotic and Prosthetic Education. (2002). Issues Affecting the Future
Demand for Orthotists and Prosthetists. Nielsen, C. C.
69
Open Bionics. (n.d.). Ada Robotic Hand - Open Bionics. Retrieved September 07, 2016, from
http://www.instructables.com/id/Ada-Robotic-Hand-Open-Bionics/
Pace, E. (2014, November 11). Vet's Self-Cooling Prosthetic Could Help Amputees Beat The
Heat. Retrieved September 07, 2016, from
http://www.npr.org/2014/11/11/363313691/vets-self-cooling-prosthetic-could-helpamputees-beat-the-heat
Pierluigi, R. (2011). On the Design of Underactuated Finger Mechanisms for Robotic Hands.
Retrieved from http://www.intechopen.com/books/advances-in-mechatronics/on-thedesign-of-underactuated-finger-mechanisms-for-robotic-hands
Radmand, A., Scheme, E., & Englehart, K. (2016). High-density force myography: A possible
alternative for upper-limb prosthetic control. Journal of Rehabilitation Research &
Development, 53(4), 443-456.
RB1. (2013, March 7). Robot Torque Arm Calculator. Retrieved October 1, 2016, from
http://www.robotshop.com/blog/en/robot-arm-torque-calculator-9712
Saladin, K. S. (2012). Anatomy & physiology: The unity of form and function (6th ed.). New
York, NY: McGraw-Hill.
Sanders, J. E., & Cassisi, D. V. (2001). Mechanical performance of inflatable inserts used in
limb prosthetics. Journal of Rehabilitation Research and Development, 38.
Seaman, J. P. (2010). Survey of Individuals Wearing Lower Limb Prostheses. Journal of
Prosthetics and Orthotics, 22(4), 257-265.
Singh, R., Hunter, J., & Philip, A. (2007). Fluid collections in amputee stumps: a common
phenomenon. Arch Phys Med Rehabil. 88, 661-663.
Street, G. (2007). Vacuum Suspension and its Effects on the Limb.
The Limbitless Arm. (n.d.). Retrieved September 7, 2016, from
http://enablingthefuture.org/upper-limb-prosthetics/the-limbitless-arm/
Andrew, Thomas J. "American Academy of Orthotists & Prosthetists." Transhumeral and Elbow
Disarticulation Anatomically Contoured Socket Considerations - Journal of Prosthetics
and Orthotics, 2008 | American Academy of Orthotists & Prosthetists. American
Academy of Orthotists and Prosthetics, n.d. Web. 13 Dec. 2016.
Understanding Thermoelectric Cooling. (n.d.). Retrieved October 08, 2016, from
http://www.activecool.com/technotes/thermoelectric.html
Williams, T. W., III. (2011). Progress on stabilizing and controlling powered upper-limb
prostheses.
Journal of Rehabilitation Research & Development, 48(6), ix-xix.
Vergara, M. et al. Journal of Hand Therapy 27 (2014) 225-234
Ventimiglia, Paul. Design of a Human Hand Prosthesis (2012): n. pag. Worcester Polytechnic
Institute, 26 Apr. 2012. Web. 11 Oct. 2016.
70
Versalius (n.d.). Retrieved from
http://www.sciencedirect.com.ezproxy.wpi.edu/science/article/pii/S0363502311009907
Zhao, D.w., L. Jiang, H. Huang, M.h. Jin, H.g. Cai, and H. Liu. "Development of a
Multi-DOF Anthropomorphic Prosthetic Hand." 2006 IEEE International Conference on
Robotics and Biomimetics (2006): n. pag. Web.
Ziegler-Graham, K., MacKenzie, E. J., Ephraim, P. L., Travison,T. G. & Brookmeyer, R. (2008).
Estimating the Prevalence of Limb Loss in the United States: 2005 to 2050. Archives of
Physical Medicine and Rehabilitation, 89, 422-429.
71
Appendices
Appendix A: Nomenclature
Term
Definition
ABS
Acrylonitrile butadiene styrene, a plastic used in additive manufacturing.
ADL
Activity of Daily Living, routine activities that people tend do every day without
needing assistance. There are six basic ADLs: eating, bathing, dressing, toileting,
walking, and continence.
CPU
Computer Fan. In most cases it is a Central Processing Unit but in this project it is used
to denote a computer fan.
EMG
Electromyography, a diagnostic procedure to assess the health of muscles and the
nerve cells that control them (motor neurons). Motor neurons transmit electrical signals
that cause muscles to contract. An EMG translates these signals into graphs, sounds or
numerical values that a specialist interprets.
MQP
Major Qualifying Project, a major-related project at Worcester Polytechnic Institute.
PLA
Polylactide, a plastic used in additive manufacturing.
72
Appendix B: Complete Directional Terms in Human Anatomy
Term
Definition
Sagittal Plane
Passes vertically through the body. Divides the body into right and left
sections
Median Plane
Sagittal plane that divides the body into two equal halves
Frontal Plane
Passes vertically through the body, but is perpendicular to the sagittal plane.
Divides the body into anterior and posterior sections
Transverse
Plane
Passes horizontally through the body. Divides the body into superior and
inferior sections
Ventral
Toward the front or belly
Dorsal
Toward the back or spine
Anterior
Toward the ventral side
Posterior
Toward the dorsal side
Supine
Anterior face up
Prone
Posterior face down
Cephalic
Toward the head or superior end
Rostral
Toward the forehead or nose
Caudal
Toward the tail or inferior end
Superior
Above
Inferior
Below
Medial
Toward the sagittal or median plane
73
Lateral
Away from the sagittal or median plane
Proximal
Closer to the point of attachment or origin
Distal
Farther from the point of attachment or origin
Ipsilateral
On the same side of the body
Contralateral
On opposite sides of the body
Superficial
Closer to the body surface
Deep
Farther from the body surface
74
Appendix C: Activities of Daily Living
Appendix C.1: Detailed Descriptions of the Activities of Daily Living
From Vergara, M. et al. Journal of Hand Therapy 27 (2014) 226
Appendix C.2: Hand Usage during Activities of Daily Living
From Vergara, M. et al. Journal of Hand Therapy 27 (2014) 228
75
Appendix D: Design Goals Decision Matrix
76
Appendix E: Schematic for Control System
77
Appendix F: Full Test Procedures
Appendix F.1: Pins Test Procedure
Background
The 3D Printed pins in the hand must support the weight of the objects the user desires to interact
with. The pins are present at the base of the fingers, connected to the palm and connecting the
fingers in the joints. Understanding how the pins will respond under different loads will determine
if the pins are ready to be used in the final assembly as they are designed. Failure to hold the
load, then a redesign will be necessary.
Abstract
Objective: The objective of this test is to determine whether the pins placed at the finger joints
will be able to withstand a tensioning force, up to 10 lb. for a 100% passing rating. All testing will
be done in a tensioning Instron machine. The classification of success and failure is described
below.
Design: The pins will be rated on percentage of desired strength to experimental values. 0% will
act as the low (failing) end of the scale while 100% acts as the highest (successful). The
breakdown of the individual ratings is below. Weight (in pounds) supported will be characterized
by ‘X’.
The pins will be rating on that scale and further categorized into either a Failure or Success.
Failure

A pin of a rating lower than 80%
Success

A pin of a rating of 80% or higher
A success signifies that no further design is needed. A failure signifies that a redesign is
needed. A video recording will be done of the test to review any failures or gap in the procedure
that could lead to failure. This will be useful in reviewing why a failure might occur and what
steps should be taken to rectify the issue or alteration in the design.
The experiment will be run with three different types of joint pins; horizontally printed, vertically
printed, and pins printed with an alteration to the flex cut. With the data, the pin type that
completes the experiment successfully or the highest rating will be pursued for the final design.
78
Experimental Procedure
Equipment List
1. Finger Assemblies, as shown in the image below
2.
3 assemblies
 One with Pins printed horizontally
 One with Pins printed vertically
 One with Pins printed with alteration to the flex cut.
3. Instron Machine.
4. Data Collection software.
5. Extra pins and finger components in case of failures.
 At least three whole Finger Assemblies
Experiment
A traditional approach to mass measurement is by utilizing a digital weighing scale.
1. The finger assembly will be placed the instron vices with the Finger Top in the upper vice
and the bottom Finger Mid in the bottom vice.
2. The vices will be tightened to a level that grips, but does not damage the components.
3. The instron will pull the finger up to 10 lb., or fracture, whichever occurs first.
4. The experimenter will record the data in both the software and the data collection table
79
(Table 2).
5. The experiment will be repeated for the 3 assemblies outlined in point 1 of the Equipment
List to determine which pin/finger assembly is strongest.
Table 2: Rating Calculation
Pin # (Type)
Joint Pin 1 (½ Cut
Design)
Joint Pin 2
(Horizontal)
Joint Pin 3
(Vertical)
Total
Load (N)
Total
Load (lbf)
Rating
(%)
250
56.2
562
400
89.9
899
60
13.5
135
Notes
Pass/Fail
Fractured right in middle
in part, image below.
Fractured at the end of
the cut, image below
Fractured at the edge of
the cut, image below
Pass
Pass
Fail
Results and Conclusions
All pins passed the test on a numerical level as they all exceeded 10 lb. by the time they
catastrophically failed. However, Pin 3 (Vertical) began fracturing before the mark was reached,
leading the experimenter to deem it as a failure. Below is an image of the results and the pins
right after their testing.
80
From this image, the location, force, and manner in which the fracture occurred can be
observed.
As the pins did exceed the expectations of the test, the test was deemed an overall
success. The pin that will be used in the final model is Pin 2 (Horizontal). This pin performed the
best of the three. The design was developed to be strong so the weakest point wouldn’t be in
the middle, where the weight would be concentrated. As the ½ Cut pin performed successfully
as well, it did not perform the best because the weakest point, where the middle cut is placed, is
where the weight was fully concentrated. Below is the image of how Pin 1 (½ Cut) fractured
within the assembly. The other pins fractured in a similar manner.
The wings of the finger do appear to be damaged in this image. However, they only
elastically deformed and returned to their original shape when released from the Instron. There
was no other damage to the finger components.
There is no need for a re-design as the pins tested past the requirement of the
prosthetic’s application.
Error Analysis
There are sources of possible error. The Instron is not expected to be inaccurate, but
calibration can always prove to be an issue. Under the supervision of a lab monitor, the
experimenter will calibrate the machine per the machine’s standard procedure. Another source
of error is pin placement. If the pin is not seated properly in the assembly, the results could be
inaccurate. Before being placed in the machine, the pin will be seated correctly by the
experimenter. Once loaded into the machine, the experimenter will visually confirm the pin is still
seated properly. If it is not, the assembly will be taken out and the pin will be reseated and the
process will repeat.
81
Comments and Reflections
There is not too much to comment on as the pins performed as expected. It does stand
out to how much weight the pins were able to support compared to what was needed of them.
However, the strength is consistent with the tensile and shear strength of PLA and the
characteristics of 3D Printing PLA. With these all considered, it is clear that the results are
consistent with how the pins were designed and manufactured. There are no suggestions for
how to improve the design as it passed the test.
82
Appendix F.2: Finger Actuation Test Procedure
Abstract
Background
Actuation of the fingers is essential to the functionality of the prosthetic design. Without
proper function the hand can cause damage to property or even injury to the user. Inconsistency
in ease of use will result in a poor product and ultimate abandonment by the user. Though not
every prosthetic arm has the ability to actuate fingers, those that do must do so correctly or the
benefits of a more expensive device are lost.
Objective
To determine if the design of the finger is able to complete the full motion required to
complete desired activities.
Description
Each trial should be recorded to determine how long it takes the finger to complete a full
closing and opening. The motion will also be recorded to determine if there is a smooth transition
between fully opened and closed, and to determine if the joints function properly. The
classification of Failure and Success are defined below.
Equipment List
1. 4 x Fully assembled fingers
2. 1 x Thumb
3. 1 x Palm
4. 10 x 6 inch lengths of thread
5. 4 x HS-5585mh servos
6. 1 x 5v power supply
7. 1 x Control Board
8. 1 x Arduino
9. 1 x 7.4V Battery
10. 1 x Servo housing
11. 10 x 8-32 ½ inch bolts
12. 10 x 8-32 nuts
Procedure
●
●
Assemble the finger(s) to the hand using the provided pins
○ The fingers must be threaded to the servo such that the finger can be actuated
properly (see assembly instructions)
Attach the servo to the power supply
83
●
●
●
●
●
Test the finger using the first, provided, prewritten program
Ensure the finger completes one full cycle of opening and closing
○ The finger should start in the open position.
The following data should be recorded:
○ Time required to complete one cycle
■ note: the time recorded should be when the signal for actuation is started
to when the finger reaches its fully closed or open position
○ Pass or fail of each cycle
○ Time required to complete (50) cycles
○ Observations of any kind, especially the smoothness of the cycle(s)
Test the finger using each of the next, provided, prewritten programs and record the
requested data above.
In order to measure continuous actuation success during testing, each actuation will be
measured using a protractor to ensure the hand is closing properly and consistently at
equal angles.
Failure
● Inability to close at all
● Inability to close fully within (5) seconds
○ Finger “closed” is defined as all three segments of fingers being used in the specific
program curling inward at least 110 degrees around their respective pin joints.
● Inability to open at all
● Inability to open fully within (5) seconds
○ Finger “open” is defined as all three sections of all four fingers lying parallel to the
bottom of the palm (angle between joints equal to or exceeding 160 degrees).
● Inability to follow the same motion for (50) cycles
● Dealignment of the finger/joints
Success
● Finger can close/open for all (50) cycles without failures
○ Examples of major failures include but are not limited to: pin separation, thread
separation, thread breaking, finger breaking
● Finger “open” is defined as all three sections of all four fingers lying parallel to the bottom
of the palm (angle between joints equal to or exceeding 160 degrees).
● Finger “closed” is defined as all three segments of fingers being used in the specific
program curling inward at least 110 degrees around their respective pin joints.
Notice
●
●
With continued use, wear of the thread and plastic components may affect the results of
these findings negatively
With continued use, particulates such as dirt, oil and water may affect the results of these
findings negatively
84
●
Failure of servo receiving or sending information for actuation does not mean the arm fails
“Finger Test- Actuation”
Data Collection and Analysis
Table 1: Test Recording
Actuation
#
Time to
Angles of Close
Actuate
(Tip, Middle,
(seconds) Bottom)
Angles of Open
(Tip, Middle,
Bottom)
Time
to
Open
Success?/Notes
1
1.8
101 108 80
170 160 80
FAIL
2
1.8
101 108 80
170 160 80
FAIL
3
1.8
101 108 80
170 160 80
FAIL
4
1.8
101 108 80
170 160 80
FAIL
5
1.8
101 108 80
170 160 80
FAIL
6
1.8
101 108 80
170 160 80
FAIL
7
1.8
101 108 80
170 160 80
FAIL
8
1.8
101 108 80
170 160 80
FAIL
9
1.8
101 108 80
170 160 80
FAIL
10
1.8
101 108 80
170 160 80
FAIL
11
1.8
101 108 80
170 160 80
FAIL
12
1.8
101 108 80
170 160 80
FAIL
13
1.8
101 108 80
170 160 80
FAIL
14
1.8
101 108 80
170 160 80
FAIL
15
1.8
101 108 80
170 160 80
FAIL
16
1.8
101 108 80
170 160 80
FAIL
17
1.8
101 108 80
170 160 80
FAIL
18
1.8
101 108 80
170 160 80
FAIL
19
1.8
101 108 80
170 160 80
FAIL
Results and Reflections:
85
The result of this testing done on February 24th 2017 was a failure, but the mechanism shows
promise. In earlier design we had overlooked the fact that our servo motors, which have a range
of 200 degrees, only have a programmable range of 110 degrees. This lack of motion illustrates
why we were unable to obtain full actuation. We believe these results to be promising because of
the repeatability of our actuations. This test was run simultaneously with the pin friction test and
showed almost identical results over 600 actuations. Moving forward we have two options to have
a successful test run. Our first option is to increase the radii of the cable tracks to compensate for
the lack of arc length. Our second option is to purchase a servo programmer to unlock the
remaining 90 degrees of motion our motors are capable of.
**Upon re-testing, the test was a success. However, the team has deemed this experiment
a failure due to concerns users may not be able to repeat the results at this time.**
Actuation #3
Actuation #300
Actuation #590
86
Appendix F.3: Thread Ductility Test Procedure
Abstract
Objective: The threads used in the prosthetic must be able to support different weights the
prosthetic will be used to pick up on a daily basis. The threads run through each finger and are
attached to servos in the wrist. When the servos rotate, the thread is pulled on and then the finger
is closed. Not only is it important to understand whether the thread can be used to actuate the
fingers, but also see if it can withstand holding weights as described in the user requirements. If
the thread deforms more than is described in the failure section then a different thread will have
to be used to actuate the fingers.
Design: To determine if the thread selected will deform while holding weights described in the
user requirements. The user requirements state the prosthetic must be able to hold 10 pounds
without elongating 3% to be considered for use. The amount of thread used in each finger is 12
inches. For the finger to actuate accurately the thread is cut to a precise length. If the thread
elongates longer than 3% then it will cause errors in the program used to actuate the finger. The
thread will have the weight attached and free hanging for 5 minutes. The classifications for
success and failure are described as:
Failure
● Greater than 3% elongation
● Breakage of a load less than 10 lb.
Success
● Less than 3% elongation
● No breakage under a load less than 10 lb.
Background
The thread must meet the requirements of being able to support at least 10 pounds while showing
minimal elongation. If the thread elongates past 3% when it is used to actuate the finger the
change in length will make it less efficient in closing the fingers. If the thread is deemed a failure
then it cannot be used for the prosthetic.
An intron will be used to test the thread. The weights that are being used will be different gallon
jugs of water. The various jugs will have weights that are either 2, 3, 5, 8, & 10 pounds. Thread
segments of six inches each will be tested. Each one of the jugs will be tied to the thread and then
allowed to hang freely. After five minutes of hanging the jug will be untied and the next jug will be
tested on a new length of thread.
Experimental Procedure
Experiment List
1. Thread segments.
2. Hanging masses
3. Ruler
87
Experiment
A traditional approach to mass measurement is by utilizing a digital weighing scale.
1. The thread will be placed in a vice with the one thread end in the upper vice and the other
thread end in the bottom vice.
2. The vices will be tightened to a level that grips, but does not damage the thread
3. The hanging masses will be tied to the finger to first for 2lb, held for 60 seconds, released,
inspected for damage.
4. If not damage, repeat for 3lb, and so on until 10lb or a failure takes place.
5. The experimenter will record the data during each weight equivalent.
6. The experiment will be repeated for a total of 3 assemblies to ensure repeatability.
The following table will be used to determine whether the thread is a success or a failure.
Weight
Used
(pounds)
Thread
Failure?
(Yes/No)
Elongation
Percentage
(%)
Success or
Failure (<
3%)
Notes
2
No
0%
Success
The thread was able to hold
the weight without any
problems
3
No
0%
Success
The thread was able to hold
the weight without any
problems
5
No
1%
Success
The thread was measured to
have elongated, but visibly
remained the same
8
No
1%
Success
The thread was measured to
have elongated and started to
become unraveled.
10
No
2%
Success
The thread was able to hold
the weight, but a constant
problem that occurs is it
unravelling after some time
88
The data will be gathered for at least three different trials. The notes section can be used to
describe if there were any abnormalities with the test procedure. An example is if the thread came
unraveled during the test, but did not break.
Results and Conclusions
The thread that was tested (Kevlar Thread) was able to support the desired weights while keeping
within the elongation percentages allowed. The most the thread elongated was while using 10
pounds. The length of thread that was used was 12 inches and the longest that it was elongated
was only .2 inches which is well within the range of passing. The Kevlar thread has been
determined to be able to support common weights that a person would encounter on a day to day
basis.
Error Analysis
There are possible places for errors to occur during the experiment. The accuracy of the
equipment being used leave room for error as they could not be precise enough to capture the
accurately determine at what weight the thread failed at. The thread length can also vary from
segment to segment. Each will be cut to as accurately as possible, however, a change in the
length can affect the results of the test. Each segment will be measured before each test to
ensure that the segments are as close to each other as possible.
Comments and Reflections
The Kevlar thread did well under the various weights it was tested with. An area of concern,
however, that has come up numerous times is with the thread unravelling. The thread is braided
together with 3 different strands and when the thread is constantly tied and untied it quickly
comes apart. The thread is still usable, but makes it difficult to string through the fingers or servo
heads when they get frayed at the ends. There has not been another thread that we have
currently found that would be able to yield the same results as the Kevlar thread so the thread
will still be used. If there was another thread that was as strong as the Kevlar, but did not come
easy unraveled that would be the ideal thread to use in the hand.
89
Appendix F.4: Socket Slippage Test Procedure
Abstract
Objective: The objective of this test is to measure the slippage of the prosthesis on the arm. To
simulate normal loading and unloading during daily use, the prosthesis will be affixed to a model
of a residual limb as if it were an actual amputee’s arm and tightened securely. Hanging masses
will be hung from the prosthesis for one minute. Multiple angles will be tested. The starting and
ending positions of the prosthesis will be marked and measured. A Likert Scale will be used to
rank the level of success or failure after the five minutes have passed.
Design: To test the slippage of the socket, the socket will be attached to the model of the residual
limb at one of three test angles: 0 degrees (horizontal), 45 degrees, and 90 degrees (vertical).
Hanging masses in the form of 2.5 lb. weights are to be hung from the fingers. Marked slip
distance after one minute of testing will be measured with calipers that have been calibrated
properly.
To test slippage on human skin, the socket portion of the prosthesis with the base plate removed
can be worn on an actual human arm. The same test masses can be hung and the test can be
repeated.
Table 1: Likert Scale for Ranking
Rank Slip Distance
7
< 0.5 cm
6
0.6 - 1.0 cm
5
1.1 - 1.5 cm
4
1.6 - 2.0 cm
3
2.1 - 2.5 cm
2
2.6 - 3.0 cm
1
> 3.0 cm
Failure
● Under any of the test mass loadings, a rating less than 5.
Success
● Under any of the test mass loadings, a rating greater than 5.
A success signifies that the current design is sufficient and no further design modifications are
needed. A failure signifies that a redesign is needed. A video recording will be made of the test
to review any failures or gaps in the procedure that could lead to failure. This will be useful in
90
reviewing why a failure might occur and what steps should be taken to rectify the issue or
alteration in the design.
Background
Due to the nature of human skin expanding and contracting with changes in internal and
external temperature, the prosthesis will inevitably shift on the user’s arm. It will also slip during
general daily use due to natural loading and unloading, sweating, and the motion of the arm.
This slippage should be limited as much as possible to ensure a fit that is both tight and
comfortable. The largest issue associated with slippage is that if the prosthesis slips too
severely it runs the risk of falling off, however slippage can also cause excess irritation to the
user. A Likert scale, shown below, is used to rate the test results.
Experimental Procedure
Equipment List
1. Entire Hand Assembly
2. 2.5 lb. weights
3. Hooked bungie cords
4. Calipers
Experiment
Before the procedure is started, four one-gallon jugs of water will be filled to 2.5, 5, 7.5, and 10
pounds of water to serve as hanging masses.
Procedure
● Attach the prosthesis to a model of a residual limb using the socket and securing straps
● Align the model to a test angle
● Test angles will be 0 degrees (vertical), 45 degrees, and 90 degrees (horizontal)
● Hang a test mass on the prosthesis
● Test masses will be 2.5, 5, 7.5, and 10 lb.
● Mark the “zero” position, where the prosthesis started before timing
● Leave the mass hanging for 5 minutes
● Measure the amount of slippage
● Repeat for two trials of each mass per test angle
● The following data will be recorded:
● Test angle
● Test mass
● Slip distance
91
Results and Conclusions
Test Angle: 0
Slippage Distance (mm)
Trial
2.5 lb.
5.0 lb.
7.5 lb.
10.0 lb.
Rank
1: Joe
0 mm
2 mm
3 mm
4 mm
7
2: Rae
1 mm
2.5 mm
5 mm
12 mm
6
3: Cameron
2 mm
3 mm
4 mm
5 mm
7
Test Angle: 45
Slippage Distance (mm)
Trial
2.5 lb.
5.0 lb.
7.5 lb.
10.0 lb.
Rank
1
0 mm
2 mm
3 mm
3.5 mm
7
2
0 mm
2.5 mm
4 mm
7.5 mm
6
3
1 mm
2 mm
2.5 mm
3.5 mm
7
Test Angle: 90
Slippage Distance (mm)
Trial
2.5 lb.
5.0 lb.
7.5 lb.
10.0 lb.
Rank
1
0 mm
0 mm
0 mm
0 mm
7
2
1 mm
3 mm
5 mm
6 mm
6
3
0 mm
0 mm
1 mm
1 mm
7
As shown in the tables, the prosthesis ranked between 6-7, meaning the test passed. In general
there seemed to be minimal motion in the prosthesis, with a lot of emphasis on shoulder
strength rather than arm strength when resisting the weights. This test should be run again with
the whole assembly, rather than just the socket, if at all possible.
Comments and Reflections
●
●
●
After the weight is released, the socket returns to its original position
Minor discomfort noted when holding 10 lb.
Feels like it is the edges are pinching the arm on the back of the forearm
92
●
●
Emphasis on shoulder muscle strength
Sock stretching more than motion of the socket
Error Analysis
The weight distribution should be more accurately regulated; the way the weights were hung, they
tended to move more towards the back.
93
Appendix F.5 Control System Hardware Test Procedure
Abstract
Objective: The objective of this test is to assess the functionality and efficiency of the hardware
portion of the hand control system which is used in hand actuation, force sensing for actuation
feedback, grip selection input and user display, and myoelectric flex sensing. Each connection on
the board will be compared to the final schematic designs for the system and a multimeter will be
used to ensure the appropriate voltages and currents are present throughout the system. Success
of the hardware control system is contingent on a pass from the software control system test.
Design: The test is designed such that each part of the control system is compared to its
respective design diagram and checked for accuracy. It can be assumed that if all parts of this
tests pass but the hand functionality test fails then the problem lies in the initial circuit design.
Background
The control system for the hand was implemented using various pieces of electronic hardware in
order to more efficiently control various aspect of the hand’s functionality. The hardware was
assembled by hand in a lab and thus must be tested to ensure all parts of the system are operating
as expected. Also inherent in electrical systems are the possibilities for inefficient power usage
and excessive heat output. The servos controlling the fingers’ motion will be tested as well as the
battery supplying power to the servos. Using various lab materials, the hardware for the control
system will be thoroughly examined and tested for full functionality.
Appendix E is a schematic for the entire system and is an integral part of ensuring the validity of
the assembled hardware. The schematic specifies the exact route of voltage between various
parts of the system and thus should be understood fully in order to perform the experimental
procedure in the next section.
The following is an outline of each component of the hardware and the expected state for each
component to constitute a viable and working system:
Myoelectric Sensors
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Signal from sensor remains relatively static (± 5 units) when electrodes are in position and
arm is at rest
Figure 1 Myoelectric Sensor Schematic Symbol
Force Sensors Array
94
●
●
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
6-pin female connector is secure and connection is being made to all 6-pin wires of ribbon
cable
An oscilloscope will be used to measure the appropriate time interval necessary for
probing each force sensor, frequency must be greater than 100Hz
Outputs from force sensors remain relatively static when no force is applied (± 5 units)
Figure 2 Force Sensor Array Schematic Symbol
Rotary Encoder Input
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Output from rotary pins matches expected output from datasheet
Figure 3 Rotary Encoder Schematic Symbol
BCD to Seven Segment Decoder & Display
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Output from pins match datasheet for each possible input signal
o Appropriate number is displayed on screen, relative to input
95
Figure 4 Decoder and Display Schematic Symbol
Servo Power and Data Headers
●
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Data input signal when no signal is applied rests within a specified voltage range (± .25V)
and thus is unaffected by noise
Servos are able to be controlled to any specific degree of rotation (118.5°) with specified
speed (079°/μsec)
Figure 5 Servo Power and Data Schematic Symbol
7.4 LiPo 2C Battery
●
●
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Battery supplies specified voltage for appropriate period of time (2200mAh)
With maximum load applied to battery, current is still supplied at the appropriate voltage
rating (7.4V ± 1V)
No schematic symbol
96
DC to DC Converter (Arduino Power Management)
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Output from converter is a steady 5V source (± 0.5V)
Figure 6 Power Supply Management Schematic Symbol
ATMega328P Power and External Clock
●
●
●
All connections match schematic, are secure, and are electrically isolated from all other
components
Arduino operates at external clock frequency (16MHz)
Arduino is supplied with output signal (5V ± 0.5V) from DC to DC converter
97
Figure 7 ATMega328P Schematic Symbol
Experimental Procedure
Equipment List
1.
2.
3.
4.
5.
6.
Entire Control System Assembly
Assembly Schematic
Power Supply capable of 9V @ 3A
Multi-Meter (V, A, Ohms) w/ Probes
Oscilloscope with four probes
Computer running Arduino IDE
Experiment
Procedure for Ensuring System Matches Schematic
● Isolate respective section of schematic diagram
● Choose wire on diagram and determine all connections to wire
o Wires with node may have more than two connections
● Use multi-meter to ensure current flow between each possible connection through
specified wire
● Repeat for all wires in section
98
General Procedure
● Myoelectric Sensor:
● Ensure connections match schematic
● Connect myoelectric sensors to Arduino
● Apply resistive pads to appropriate locations on skin
● Connect sensor to pads
● Use oscilloscope to look at data signal with no flex
● Measure variation in signal and compare to expected signal as per the datasheet
● Force Sensor Array:
● Ensure connections match schematic
● Connect force sensors to 6-pin male connector
● Use multimeter on resistance setting across male to female connection
● Infinite reading means failure
● General resistivity of wire means success
● Attach oscilloscope lead to each pin of shift register and measure timing of signal
● Calculate frequency of complete measurement cycle
● Compare calculated frequency to expected frequency (100Hz)
● Attach oscilloscope lead to analog output pin
● Measure variation in signal and compare to expected signal variation (± 0.25V)
● Rotary Encoder Input
● Ensure connections match schematic
● Attach oscilloscope lead to each output pin of encoder
● Verify that output matches appropriate value as outlined in datasheet
● BCD to Seven Segment Decoder & Display
● Ensure connections match schematic
● Attach oscilloscope to each output of decoder
● Verify that output matches appropriate value as outlined in datasheet
● Use multimeter to probe between display and output of decoder
● Determine if all pins are effectively connected
● Servo Power and Data Headers
● Ensure connections match schematic
● Attach oscilloscope to each data input
● Measure variation in signal and compare to expected signal variation (± 0.25V)
● Use Arduino software to load a sweep program on Arduino
● Attach each servo separately to power
● Verify that 200 degrees of rotation is achieved
● 7.4 LiPo 2C Battery
● Ensure connections match schematic
● Use multimeter to measure voltage across battery
● Connect one servo to battery
● Use sweep servo program to continuously run servo
● Measure current and voltage
● Compare values to specification
● Connect four servos to battery
● Use sweep servo program to run all servos continuously
● Verify that battery output remains sufficient
● DC to DC Converter (Hardware Iteration #2)
● Ensure connections match schematic
● Use multimeter to measure input voltage and current
● Use multimeter to measure output voltage and current
99
●
●
Apply max load to converter
● Force sensors are being polled
● Myoelectric sensors are being polled
● Grip Selection is active
● Servos are active
● Ensure output remains constant and unchanging (± 0.25V)
ATMega328P Power and External Clock (Hardware Iteration #2)
● Ensure connections match schematic
● Determine clock speed of Arduino using oscilloscope
Resources Necessary for Testing
Myoelectric Sensor:
● Datasheet
o https://cdn.sparkfun.com/datasheets/Sensors/Biometric/MyowareUserManualAT04001.pdfhttps://cdn.sparkfun.com/datasheets/Sensors/Biometric/MyowareUserMa
nualAT-04-001.pdf
https://cdn.sparkfun.com/datasheets/Sensors/Biometric/MyowareUserManualAT-04-001.pdf
74HC595 Shift Register:
● Datasheet
o https://www.sparkfun.com/datasheets/IC/SN74HC595.pdfhttps://www.sparkfun.c
om/datasheets/IC/SN74HC595.pdf
https://www.sparkfun.com/datasheets/IC/SN74HC595.pdf
Rotary Encoder:
● Datasheet
o https://www.sparkfun.com/datasheets/Components/TW700198.pdfhttps://www.sparkfun.com/datasheets/Components/TW-700198.pdf
https://www.sparkfun.com/datasheets/Components/TW-700198.pdf
CD4511B Decoder:
● Datasheet
o http://www.ti.com/lit/ds/symlink/cd4511b.pdfhttp://www.ti.com/lit/ds/symlink/cd451
1b.pdf
http://www.ti.com/lit/ds/symlink/cd4511b.pdf
HS-5585MH Servo:
● Specification
o https://www.servocity.com/hs-5585mh-servohttps://www.servocity.com/hs5585mh-servo
https://www.servocity.com/hs-5585mh-servo
Results and Conclusions
Portion of Test
Schematic
Match?
Measurement
Notes
Result
100
Myoelectric
Sensor
Yes
Force Sensor
Array
No (See
notes)
Min: 575mV
Max: 730mV
Signal Wire
Variation
No specified
noise ratio in
datasheet; able
to differentiate
flexed from
relaxed
PASS
QD(pin 3)
inf resistivity
QG(pin 6)
inf resistivity
2 force sensors
unable to send
data
FAIL
Shift register frequency:
100Hz
Rotary Encoder
Input
Yes
Matches 100Hz
frequency
PASS
Min: -30mV
Max: 50mV
Refers to
variation in
analog
signal
<.25V noise
difference
PASS
1101, 0100, 0010, 1011 > clockwise
1110, 0111, 0001, 1000 > counter-clockwise
Used
software to
determine
pin values
Recorded bit
values match
datasheet
PASS
Check
datasheet
BCD to Seven
Segment
Decoder &
Display
Yes
Output matches
datasheet
Decoder output
matches
datasheet
PASS
Correct number
seen on display
PASS
Servo Power and Yes
Data Headers
min, max for each servo:
-28mV, 56mV; -28mV,
56mV, -28mV, 56mV; 28mV, 56mV
<.25V noise
difference
PASS
101
writeMicroseconds: 750
to 2245 -> ~160 degrees
rotation
7.4 LiPo 2C
Battery
DC to DC
Converter
(Arduino Power
Management)
second
iteration;
unable to
test
ATMega328P
Power and
External Clock
second
iteration;
unable to
test
able to move
minimum
required distance
and speed
PASS
7.62V,
One servo
After running
sweep program,
sufficient power
output for one
servo
PASS
7.58V
Four servos
After running
sweep program,
sufficient power
output for four
servos
PASS
Error Analysis
The only failure was infinite resistivity in pins 3 and 6 of the force sensor array, which
signifies that the connection between the shift register and the pin is broken. This results in the
respective force sensor for each not being able to be read by the microcontroller. The solution
for this is to check the connection of each pin and rewire and/or re-solder each as necessary.
Conclusion
The tests were overall successful. All results of the tests were within or surpassed the
specifications. The only issue experienced was two of the force sensor pins were disconnected
with the shift register. The ATMega328P Power and External Clock was unable to be tested as
that is a component of a custom PCB board that, while designed, is not physically created as of
the time of writing. This also applies to the DC to DC Converter. Once the second iteration of
board is created, these aspects will be tested.
102
Appendix F.6 Hand Functionality Test Procedure
Abstract
Objective: The objective of this test is to assess the functionality of the prosthesis’ preprogrammed grips and gestures. These are: Open Hand, Closed Hand, Point, and Pinch. To do
this, a scale of 1-7 is used to determine how well the prosthesis accomplished an assigned task,
with 7 being success and 1 being failure.
Design: For this test, the prosthesis should be either secured to a test subject’s arm (note: the
base plate should be removed if the subject does not have a residual limb), or to a model of a
residual limb. Once secure, the prosthesis will be tested using the procedures and objects
outlined below. Each grip/gesture is to be tested with different objects as appropriate.
Failure
● Inability to complete the assigned task.
Success
● Ability to complete the assigned task.
A success signifies that the current design is sufficient and no further design modifications are
needed. A failure signifies that a redesign is needed. A video recording will be made of the test
to review any failures or gaps in the procedure that could lead to failure. This will be useful in
reviewing why a failure might occur and what steps should be taken to rectify the issue or
alteration in the design. Each grip will either succeed or fail.
Background
The prosthesis needs to be able to function as a system, not just as individual components. To
this end, this test aims to assess how well the prosthesis can accomplish given tasks. This serves
as an adequate measure of how functional the system is.
Experimental Procedure
Equipment List
1. Entire Hand Assembly
2.
3.
4.
5.
6.
7.
7.4V Power Supply
Key
Tennis ball
Gallon jug
Soda Can
Door knob
Experiment
Each grip will have its own specific test and procedure to follow. The grips to be tested are:
Open Hand, Closed Hand, Point, and Pinch. Each grip will be tested twice.
Initial Setup
● Assemble the prosthesis with the appropriate force sensors and electronics in place
o Force sensors should be placed in 12 locations on the hand and fingers
o Kevlar thread should be strung through the fingers and palm using the appropriate
thread channels and holes, and attached to the servos located in the forearm
103
●
●
Test the hand to determine if the fingers can move using the muscle sensors
o Force sensors should be able to identify when they are in contact with an object
o Muscle sensors should be placed on the arm in their appropriate, predetermined
locations
▪ The forearm’s flexor and extensor muscles are used to open and close the
fingers, respectively
o Each finger should be tested individually as necessary
Test the control system by adjusting the knob
o observe whether or not each grip/gesture can be properly selected and executed
Open Hand Test
● All grips should default to open hand initially
o Ensure that each grip/gesture toggles appropriately between the gesture and all
fingers open by flexing and extending
Closed Hand Test
● Flex, with no object blocking the fingers. Ensure the hand can close into a fist, and then
reopen
● Test the ability of the fingers to close around objects of various shapes and sizes
o Specifically test: Doorknob, gallon jug handle, soda can, tennis ball
▪ Each of these objects offers a unique shape that the fingers should be able
to close around.
● Record results and observations
Point Test
● Ensure that the index finger does not actuate during the opening and closing of the hand
into this gesture
Pinch Test
● Ensure the ring and little fingers do not interfere with the closing of the thumb, index, and
middle fingers into the pinch grip.
o These fingers should close first, followed by the pinching fingers
● Determine if the pinching fingers can adequately grip a key
o Test if the key can be inserted into a lock, record observations
Comments and Reflections
Results and Conclusions
This test has been deemed a failure due to the actuations test failure. The experimenters
believe the test would succeed but are not entirely confident that other experimenters and users
can reliably repeat a successful actuation test, rendering functionality a failure.
Error Analysis
To compensate for possible error in either the 3D printed components of the hand or the
driving control system, each grip or gesture will be tested twice. However, there is still the
possibility of inconsistency due to noise, position of the object the hand is gripping, friction in the
hand, or changes in the object. This inconsistency will be controlled and limited as much as
possible.
104
Appendix G: Model Prototype Images
Appendix G.1 Bottom View
105
Appendix G.2 Isometric View
106
Appendix G.3 Manufactured Design
107