GPS BASED AUTONOMOUS CAR NAVIGATION SYSTEM Minhas Thaker B.E., Gujarat University, India, 2008 PROJECT Submitted in partial satisfaction of the requirements for the degree of MASTER OF SCIENCE in ELECTRICAL AND ELECTRONIC ENGINEERING at CALIFORNIA STATE UNIVERSITY, SACRAMENTO FALL 2010 GPS BASED AUTONOMOUS CAR NAVIGATION SYSTEM A Project by Minhas Thaker Approved by: ________________________________, Committee Chair Jing Pang, Ph.D. ________________________________, Second Reader Preetham Kumar, Ph.D. _________________________ Date ii Student: Minhas Thaker I certify that this student has met the requirements for format contained in the University format manual and that this project is suitable for shelving in the Library and credit is to be awarded for the Project __________________________, Graduate Coordinator Preetham Kumar, Ph.D. Department of Electrical and Electronic Engineering iii ___________________ Date Abstract of GPS BASED AUTONOMOUS CAR NAVIGATION SYSTEM by Minhas Thaker Autonomous car navigation system based on GPS (Global Positioning System) is a new and promising technology, which uses real time geographical data received from several GPS satellites such as longitude, latitude, speed and course to help navigate a car. The goal of the project is to make a auto-navigational car model that can route through known or pre-programmed co-ordinates autonomously without any human involvement. The project discusses how GPS readings of the current and destination points are used to compute the distance and direction of the destination and thereby navigating the car on the set path. It also discusses how the car must maintain its direction by automatically correcting its course based on new GPS data received. The project discusses various issues that were encountered and solved throughout the course of the project. The project is coded in C, developed and compiled in AVR studio software and is implemented on a Atmel ATmega328 microcontroller. The project utilizes EM408 SiRF Star III/LP single based chipset GPS engine board receiver manufactured by GlobalSat Technology Corp., Taiwan and sold by USGlobalSat Inc, USA. The output of the GPS iv receiver is a standard NMEA signal which is decoded by the microcontroller to get necessary geographical parameters. Once the microcontroller has the required data, it can compute the direction of movement and thereby navigate the car. Also, the inherent logic steers the car in case the car deviates more than a certain degree from its course. The car that is referred to in this project is 4WD dc-motor controlled robot car. __________________________, Committee Chair Jing Pang, Ph.D. __________________________ Date v ACKNOWLEDGMENTS I would like to present deep and sincerest gratitude to Dr. Jing Pang for providing me with an opportunity to work on this project. This project has immensely helped me to get exposure to the emerging technology of GPS navigation. I thank her for providing all the necessary resources, help and support for the project. I would also like to thank Dr. Preetham Kumar for reviewing my project report and providing timely feedback and valuable suggestions to improve my report. I am truly thankful to the faculty and staff members of the Department of Electrical and Electronics Engineering at California State University, who have been instrumental directly as well as indirectly in making this project successful. Lastly, I would like to thank my family and friends for providing immense support, strength and motivation throughout the course of my project. Minhas Thaker vi TABLE OF CONTENTS Page Acknowledgments.............................................................................................................. vi List of tables ....................................................................................................................... ix List of figures ...................................................................................................................... x Chapter 1. INTRODUCTION .......................................................................................................... 1 1.1. Introduction to GPS Based Autonomous Navigation .............................................. 1 1.2. Purpose of the Project .............................................................................................. 1 1.3. Significance of the Project ....................................................................................... 2 1.4. Organization of the Report....................................................................................... 3 2. INTRODUCTION OF GPS ............................................................................................ 4 2.1. GPS Overview ......................................................................................................... 4 2.2. GPS Satellite Constellations .................................................................................... 6 2.3. GPS Working ........................................................................................................... 8 2.4. Reference Surfaces and the Geodetic Coordinate System ..................................... 11 2.5. NMEA Protocol ..................................................................................................... 14 3. WAYPOINT NAVIGATION ....................................................................................... 17 3.1. Introduction to Waypoint Navigation .................................................................... 17 3.2. GPS Noise .............................................................................................................. 17 3.3. EM408 Pin Configuration and Connection with ATmega328 .............................. 17 3.4. Analysis of GPS Readings ..................................................................................... 19 4. ATMEGA328 MICROCONTROLLER ....................................................................... 24 4.1. Introduction to Atmega328 .................................................................................... 24 vii 4.2. Approach ................................................................................................................ 25 4.3. Algorithm ............................................................................................................... 27 4.4. Computing Distance and Direction of Destination Point ...................................... 30 5. EXPERIMENTAL RESULTS...................................................................................... 33 5.1. Plotting the Waypoints on a Map........................................................................... 33 5.2. Plotting the Path Traversed By the Car.................................................................. 34 6. CONCLUSION ............................................................................................................. 36 Bibliography ..................................................................................................................... 37 viii LIST OF TABLES Page 1. Table 2.2.1 GPS satellite constellation ........................................................................... 7 2. Table 2.4.1 Reference systems and associated ellipsoids ............................................. 12 3. Table 2.5.1 General NMEA sentences .......................................................................... 15 4. Table 2.5.2 RMC sentence format ................................................................................ 16 5. Table 3.4.1 Tabulating mean and standard deviation for GPS readings ....................... 22 ix LIST OF FIGURES Page 1. Figure 2.1.1 GPS segments ............................................................................................. 6 2. Figure 2.3.1 Determining position of receiver by triangulation of three satellites ....... 10 3. Figure 2.4.1 Geodetic latitude and longitude ................................................................ 13 4. Figure 2.4.2 Geodetic height ......................................................................................... 14 5. Figure 3.3.1 EM408 GPS receiver ................................................................................ 18 6. Figure 3.3.2 Pin connection of ATmega328 with GPS receiver ................................... 19 7. Figure 3.4.1 Bar-plot of standard deviation for latitude and longitude readings .......... 21 8. Figure 3.4.2 Mappings Points A, B, C and D on Google Maps .................................... 21 9. Figure 3.4.3 Gaussian distribution of latitude and longitude readings ......................... 23 10. Figure 4.1.1 Pin diagram of ATmega328 DIP ............................................................ 24 11. Figure 4.1.2 Internal architecture of ATmega328 microcontroller ............................. 25 12. Figure 4.2.1 Car mounted with GPS receiver and microcontroller ............................. 26 13. Figure 4.3.1 Algorithm ................................................................................................ 29 14. Figure 4.4.1 Calculating deviation from actual course ............................................... 32 15. Figure 5.1.1 Plot of point A, B, C and D on a map ..................................................... 33 16. Figure 5.2.1 Plot of actual path vs. path traversed by the car ..................................... 35 x 1 Chapter 1 INTRODUCTION 1.1. Introduction to GPS Based Autonomous Navigation Autonomous navigation broadly refers to any technique, approach or method, which can be utilized to safely navigate a vehicle (terrestrial, marine, airborne or deep space), on its own in a static or dynamic environment without any intervention by a human controller. Autonomous navigation is a way to help achieve better route planning, path prediction, smoother maneuverability in dynamic environments and thereby achieving optimized fuel efficiency and enhanced human comfort. The approach of GPS based autonomous navigation utilizes a GPS receiver that receives signals from a constellation of GPS satellites. The receiver then computes its position on the earth surface. A navigation algorithm then computes other parameters such as direction and distances that aid in autonomous navigation. 1.2. Purpose of the Project The purpose of the project is to develop an algorithm that takes in inputs from a GPS receiver and using those inputs to successfully navigate a car through a set of known points also known as waypoints. This type of navigation is also called terrestrial waypoint navigation or in general, waypoint navigation [1]. Waypoint navigation is also used in marine navigation which is also based on the same concept. The car described here is a 4WD dc-motor controlled robot car. In general it would be referred as a car, not to be confused with a real passenger car. The overall process is divided into three parts. In the first part, when the car is at the start waypoint the GPS receiver tries to get a fix on 2 visible GPS satellites and compute current latitude and longitude. In the second part, the algorithm in the microcontroller computes the direction of the destination waypoint from the current waypoint and drives the car. In the third step, the algorithm steers the car in case it deviates from its main course. These steps repeat itself after reaching the destination and navigates to next waypoint. 1.3. Significance of the Project GPS based autonomous navigation is a very rapidly developing technology. Researchers have developed several techniques for navigation under a variety of external environments. Such techniques utilizes robots that have a variety of sensors to sense external environment such as proximity sensors, ultrasound sensors, and a camera as shown in [2]. This project utilizes a GPS receiver to help navigate through fixed waypoints. Autonomous navigation was implemented in the mars exploration rover. Therein they implemented features such as tracking ground features, terrain mapping and obstacle avoidance, mentioned in [3]. Autonomous navigation also finds its application in defense research. This project focuses on researching a technique of terrestrial navigation as to learn about GPS receivers, its signals, understanding how such data can help to navigate a car, and then interface all components together. It will help in understanding general GPS satellite system and standard GPS protocols. Developing such a technique is useful towards researching new domains of navigation applications. 3 1.4. Organization of the Report Chapter 2 focuses on basic introduction of GPS, its general understanding, GPS satellite constellations, and standard GPS navigation. It describes how latitude and longitude coordinates are computed by a GPS receiver. It also describes general geodetic coordinate system and NMEA protocol, a standard for GPS receiver output messages. Chapter 3 gives overview about the waypoint navigation, GPS receiver used in the project, its configuration and wiring with ATmega328 microcontroller. It explains how GPS coordinate readings were taken and statistically evaluated for precision measurement. Chapter 4 talks about the ATmega328 controller, design steps, and the algorithm developed for navigation. Chapter 5 describes the experimental result; it displays output results from the GPS receiver and is plotted in a map showing waypoints. . A map of proposed path and actual path are plotted and how to smooth out car steering control is discussed. Chapter 6 focuses on the conclusion and future work 4 Chapter 2 INTRODUCTION OF GPS 2.1. GPS Overview GPS, the Global Positioning System, is a constellation of satellites that provides a user with an accurate position on the surface of the earth. This satellite based navigation system was developed by the U.S. Department of Defense (DoD) in early 1970s. It was first intended for military use but later it was made available to civilian users. GPS can provide precise position and time information to a user anywhere in the world. It is a one way system i.e. a user can only receive signals but cannot send signals to the satellite. This type of configuration is needed due to security reasons as wells as to serve unlimited number of users [1]. The GPS system consists of three main segments: 1. Space segment 2. Control segment 3. User segment The space segment consists of a constellation of 24 satellites in fixed orbits around the earth. Each satellite continuously transmits GPS signals to the earth, which consist of two carrier frequencies, digital codes, and navigation message. The carrier frequency and codes are used in determining distance of the satellite from the receiver. The navigation message consists of information like satellite location and clock compensation [1]. The control segment includes a network of tracking stations. These tracking stations continuously monitor the satellite orbit; checks satellite clock, atmospheric 5 conditions, satellite almanac, provide necessary compensation for clock error, and upload data to satellites. The OCS (Operational Control System) consists of one master control station (MCS), several monitor stations, and ground control stations. The MCS is located in the United States at Shriever Air Force Base, Colorado Springs, Colorado [1] [4]. The user segment consists of all civilian and military users. A GPS receiver can lock on to any visible satellites and can determine its location on earth’s surface. GPS system was first conceived for military usage such as navigating in remote terrains and coordinating military activities. With addition of civilian access to GPS signals, it is used in various applications such as surveying, mapping, hiking, and vehicular navigation [4]. Figure 2.1.1 shows the three segments of a GPS system. The GPS satellites transmit signals at two carrier frequencies. One of them is called L1 carrier which generates carrier frequencies at 1575.42 MHz, while the other is called L2 carrier which generates frequencies at 1227.60 MHz [4]. The master control station uploads data to the satellites via ground using frequencies in S-band (2 GHz to 4 GHz). 6 Figure 2.1.1 GPS segments [1] 2.2. GPS Satellite Constellations At present (as of October 2010) there are total of 32 satellites orbiting around earth in circular orbits [5]. The satellites are placed in six orbits labeled A to F, inclined at 55◦ with respect to the equator and at a height of approximately 20,200 km [4]. Nominally there are 24 satellites in GPS constellation, thus having 4 satellites in each orbit. Since there are more than 24 satellites, an orbit can contain more than 4 satellites. The control sites are located at various places all over the globe to monitor and control the path of satellites. The GPS satellites are categorized into six sequences of Block I, II, IIA, IIR, II-M and IIF. Each satellite is identified by its unique SVN (Space Vehicle 7 Number) [1]. Table 2.2.1 shows the sequence number, SVN and the orbital plane of current GPS satellites. Orbital Sequence SVN Orbital Plane Sequence SVN Plane II-10 23 E5 IIR-7 54 E4 II-11 24 D2-F IIR-8 56 B1 II-14 26 F2-F IIR-9 45 D3 II-15 27 A6 IIR-10 47 E2 II-21 39 A1 IIR-11 59 C3 II-23 34 D4 IIR-12 60 F4 II-24 36 C5 IIR-12 61 D1 II-25 33 C2 IIR-M-1 53 C4 II-26 40 E6 IIR-M-2 52 A2 II-27 30 B2-1 IIR-M-3 58 B4 II-28 38 A3 IIR-M-4 55 F2 IIR-2 43 F3 IIR-M-5 57 C1 IIR-3 46 D2 IIR-M-6 48 A4 IIR-4 51 E1 IIR-M-7 49 B1-F IIR-5 44 B3 IIR-M-8 50 E3 IIR-6 41 F1 IIF-1 62 B2 Table 2.2.1 GPS satellite constellation [5] 8 2.3. GPS Working When a GPS receiver locks on to a satellite it will perform three basic steps: 1. Measure distance from the satellite to the receiver. 2. Determine the position of the receiver by triangulation. 3. Synchronize GPS receiver clock to the satellite. The GPS receiver will measure the time taken by a signal from a satellite to reach the receiver. It is assumed that the receiver clock is synchronized with the satellite clock at this time. The time is then multiplied with the velocity of light c, to get the distance between the satellite and the receiver, shown in equation 2.3.1. Getting precise time measurement is achieved by a method of pseudo random code. Pseudo random code is a complex sequence of ‘on’, ‘off’ pulses. Each satellite has its own pseudo random code, hence the receiver cannot accidently pick up signals from a different satellite and all satellites can operate on the same frequency. The method works as follows. Both the satellite and the receiver will start generating pseudo random code at the same time. The satellite signal will carry this code in sequence as a part of the GPS signal. The receiver then computes how late a particular pseudo code was received compared to the code generated by it. This is how the GPS receiver determines the time taken by a signal to travel from satellite to receiver [1]. 9 d = c × t ……..…………………….……………………………………….. (2.3.1) where d is distance of receiver from satellite, c is velocity of light, 299792458 m/s, t is the time taken for signal from satellite to reach receiver. The receiver computes its distance (r1) from one satellite. With measurement from one satellite, the receiver can be anywhere on a sphere with its center as a satellite and radius r1 as shown in figure 2.3.1a. Further on, taking another distance measurement (r2) from the second satellite, the receiver can be on the intersection of two spheres as shown in figure 2.3.1b. On taking the third distance measurement (r3) from the third satellite, the receiver precisely lies on two points, which are the intersection points of three spheres as shown in figure 2.3.1c. One of the two points will lie out farther in space compared to the other; hence the former can be rejected. The point left is the position of the receiver on the earth surface [6]. For the above two steps it is assumed that the receiver clock is in synchronization with the satellite clock. The satellite clocks are atomic clocks hence they are very precise, but they also tend to give rise to timing errors. Control centers continuously monitor the satellite clocks and send correcting signals from time to time. The satellite in turn sends these corrections to the receiver as a part of its signal to synchronize the receiver clock and correct the distance measured. Also, it is impossible to have a precise clock at the receiver end and will give rise to receiver clock error. As a result, the spheres in figure 1c would not intersect at one point. A fourth parameter dT (receiver clock error) has to be included in the equation to calculate the distances. To solve four unknowns, four 10 equations are required; hence distance measurement from one more satellite is needed as shown in equations 2.3.2 to 2.3.5 On solving this equation for X, Y, Z and dT, coordinates of the receiver on the earth surface can be found [6]. Figure 2.3.1 Determining position of receiver by triangulation of three satellites [6] 11 r1= ( X − x1 ) 2 + (Y − y1 ) 2 + ( Z − z1 ) 2 − cdT ………..………....…………… (2.3.2) r2= ( X − x2 ) 2 + (Y − y2 ) 2 + ( Z − z2 ) 2 − cdT .............................................(2.3.3) r3= ( X − x3 ) 2 + (Y − y3 ) 2 + ( Z − z3 ) 2 − cdT ………….…………………… (2.3.4) r4= ( X − x4 ) 2 + (Y − y4 ) 2 + ( Z − z4 ) 2 − cdT ………………………………. (2.3.5) where r1,r2,r3,r4 are distances of receiver from satellite 1, 2, 3 and 4 respectively, X,Y,Z are three dimensional Cartesian coordinates of receiver, (x1,y1,z1), (x2,y2,z2), (x3,y3,z3) and (x4,y4,z4) are the coordinates of satellites 1,2,3 and 4 respectively, transmitted as part of GPS signal, c is velocity of light, 299792458 m/s, dT is the receiver clock error. 2.4. Reference Surfaces and the Geodetic Coordinate System Earth’s topographic surface is highly uneven, which makes accurate measurements of geodetic data highly error prone. Geodesists have formulated two reference systems that approximate the surface of the earth. These are called Geoid and Ellipsoid. A Geoid is a surface at mean sea-level used to represent height on a map. An ellipsoid is a mathematical approximation of earth surface. WGS-84, NAD 83 and GRS 80 are some common reference ellipsoid used [1] [7]. WGS-84 is an acronym of World Geodetic System of 1984. It is a reference ellipsoid used by the U.S. Department of Defense. The National Geospatial Intelligence Agency maintains and constantly enhances the WGS-84 reference standard. NAD 83 which stands for North American Datum for 12 1983 is another reference ellipsoid used mainly in North America. It is based on the ellipsoid of Geocentric Reference System of 1980 (GRS 80). Table 2.4.1 shows the common reference system used and the ellipsoid values. Reference System Ellipsoid a(m) 1/f WGS 84 WGS 84 6378137.0 298.257223563 NAD 83 GRS 80 6378137.0 298.257222101 NAD 27 Clarke 1866 6378206.4 294.9786982 Table 2.4.1 Reference systems and associated ellipsoids [1] Geodetic Coordinate system is a three dimensional coordinate system in which a point on earth’s surface is represented by three quantities latitude (ϕ), longitude (λ) and geodetic height (h). The coordinate system has three axis, the z axis coincides with rotational axis of the ellipsoid, x and y axis lie on the equatorial plane. Latitude of a point is defined as an angle between the normal drawn at the ellipsoid and the equatorial plane. Longitude of a point is the angle between the Prime Meridian which passes through Greenwich and the meridian ellipse that contains the point. Geodetic height is the distance between the point on the earth surface and the point on the ellipsoid along the line perpendicular to the ellipsoid [1] [7]. As shown in figure 2.4.1, consider a point P on the surface of the earth. The earth surface is approximated as an ellipsoid. Point P’ is a point on the ellipsoid exactly below point P. A normal to the ellipsoid is drawn from point P and intersects the equatorial plane. The angle made by the normal to the plane is the latitude (ϕ) of the point. The meridian, at point P, makes an angle (λ) with the Greenwich meridian on the equatorial 13 plane. This angle is the latitude of point P. Now considering figure 2.4.2, the distance between point P and P’ gives the geodetic height of the point. There are also other coordinate systems to determine location of a point on earth’s surface. A coordinate system can be converted to other using mathematical equations. These coordinates can also be projected on a map using various projection techniques [7]. Figure 2.4.1 Geodetic latitude and longitude [7] 14 Figure 2.4.2 Geodetic height [8] 2.5. NMEA Protocol National Marine Electronics Association is a group of electronics dealers, which has standardized an interface protocol for electrical signaling and data transmission in marine electronic equipments. This project uses NMEA 0183 format of GPS receiver output. A common GPS receiver outputs data streams based on NMEA 0183 format. Each NMEA sentence starts with a “$” sign and ends with a carriage return-line feed <CR><LF>. The “$” is followed by a five character address field (e.g. $GPRMC), of which the first two characters define the talker and the last three determine NMEA sentence. There are 8 basic types of NMEA output messages or sentences [9]. Table 2.5.1 shows some general NMEA sentences. 15 NMEA sentence Description GGA GPS fixed data GLL Geographic position - Latitude/Longitude GSA GNSS DOP and active satellites GSV GNSS satellites in view MSS MSK receiver signal RMC Recommended minimum specific GNSS data VTG Course over ground and ground speed ZDA SiRF timing message Table 2.5.1 General NMEA sentences [9] 16 This project utilizes RMC NMEA sentence. Table 2.5.2 shows detailed description of the RMC sentence. Name Example Message ID $GPRMC RMC protocol header UTC time 161229.487 hhmmss.sss Status A Latitude 3723.2475 ddmm.mmmm N/S indicator N N= north or S = south Longitude 12158.3416 ddmm.mmmm E/W indicator E E = east or W = west Speed over ground 0.13 Knots 309.62 Degrees Course over ground Date V = not valid data A Checksum *10 True Ddmmyy Degrees Mode Description A = valid data 120509 Magnetic Variation <CR><LF> Units E = east or W = west A=autonomous, D=DGPS, E=DR End of message termination Table 2.5.2 RMC sentence format [9] 17 Chapter 3 WAYPOINT NAVIGATION 3.1. Introduction to Waypoint Navigation A waypoint in general is a coordinate that identifies a point in 3-dimensional space. For terrestrial application it is generally defined using latitude and longitude coordinates. For aviation applications it is defined using latitude, longitude and altitude. Waypoint navigation is a technique by which a user can navigate from one waypoint to another. Generally, a GPS receiver is used to compute the direction of destination waypoint from current position and help navigate the user. Waypoint navigation is generally used by surveyors to layout points and lines and mariners to navigate in open sea [1] [10]. 3.2. GPS Noise The GPS signal is constantly affected with errors and noise. These errors and noise can affect the computed coordinates by the GPS receiver. Noise and error signal can creep into the GPS signal at any level starting from the satellite to the receiver. Ionosphere and Troposphere refraction at signal propagation level, and clock bias and multipath at the receiver level account for major error in GPS data [4]. 3.3. EM408 Pin Configuration and Connection with ATmega328 This project utilizes EM408 receivers manufactured by GlobalSat Technology Corporation. Figure 3.3.1 shows the EM408 receiver. EM408 GPS engine board is a SiRF Star III chipset based receiver, supports WAAS, EGNOS and MSAS augmentation, features an integrated patch antenna and supports NMEA 0183 data protocol. 18 Pin explanation: 1. Enable/Disable: On/Off. 2. VCC: 3.3 DC supply 3. TX: Transmit channel output. 4. RX: Receive channel input for receiving software commands. 5. GND: Provides ground to the engine board. Figure 3.3.2 shows connection of GPS receiver to ATmega328 board. Enable, RX and TX pins were connected to ATmega328 pins 2, 3 and 4 respectively. GND and VCC pins were connected to GND and VCC (3.3) pins of ATmega328. Figure 3.3.1 EM408 GPS receiver 19 Figure 3.3.2 Pin connection of ATmega328 with GPS receiver 3.4. Analysis of GPS Readings Several GPS readings were taken at point A, B, C and D, located in the facultystaff parking lot, at an interval of one hour. Thereafter, mean and standard deviation were computed to determine precision of GPS readings. The results obtained are shown in table 3.4.1. A bar graph showing standard deviation in meters is shown in figure 3.4.1. Mean latitude and longitude coordinates of points A, B, C and D are plotted on a map as 20 shown in figure 3.4.2. Gaussian response of latitude and longitudes readings for all points is shown in figure 3.4.3, using the mean and standard deviation from table 3.4.1. The latitude and longitude readings were fairly precise with a standard deviation ranging from 0.92 meter to 2.26 meters. For latitude readings, Point B had the lowest standard deviation while point D had the highest standard deviation. For longitude readings, Point C had the lowest standard deviation while point A had the highest standard deviation. Overall longitude readings showed less deviation compared to latitude readings. Equation 3.4.1 was used for converting standard deviation in degrees to standard deviation in meter. 1'' of φd = 30 m 1'' of λd = 30*cos (φ ) m ……………………...……………………………… (3.4.1) where ϕd is latitude standard deviation, λd is longitude standard deviation, ϕ is mean latitude for the point, 21 Figure 3.4.1 Bar-plot of standard deviation for latitude and longitude readings Figure 3.4.2 Mappings Points A, B, C and D on Google Maps 22 Latitude(ϕ) Longitude(λ) Mean 38.5603182128 -121.4210546807 Standard Deviation (deg) 0.0000104313 0.0000158458 Standard Deviation (m) 1.1265828691 1.3381948275 Mean 38.5603597584 -121.4208304601 Standard Deviation (deg) 0.0000085694 0.0000140004 Standard Deviation (m) 0.9254949024 1.1823422272 Mean 38.5607323646 -121.4210119247 Standard Deviation (deg) 0.0000125592 0.0000125281 Standard Deviation (m) 1.3563896801 1.0580017803 Mean 38.5606722938 -121.4211874816 Standard Deviation (deg) 0.0000210007 0.0000154879 Standard Deviation (m) 2.2680759935 1.3079614379 Point A Point B Point C Point D Table 3.4.1 Tabulating mean and standard deviation for GPS readings 23 Figure 3.4.3 Gaussian distribution of latitude and longitude readings 24 Chapter 4 ATMEGA328 MICROCONTROLLER 4.1. Introduction to Atmega328 Atmega328 is a high performace, low power, 8-bit CMOS microcontroller, manufactured by Atmel. It is based on modified Harvard AVR RISC architecture. It features 32K bytes of Flash, 1K bytes of EEPROM, and 2K bytes of RAM memory. It also has an optional boot code section where in a bootloader program can be loaded. Hence the feature of in-system programming can also be attained. Atmega328 is generally programmed through AVR studio. AVR studio is the Integrated Deeopment Environemt (IDE) for Atmel microcontrollers [11]. Figure 4.1.1 shows the pin diagram of ATmega328 microcontroller. The AVR RISC architecture is shown in figure 4.1.2. Figure 4.1.1 Pin diagram of ATmega328 DIP [11] 25 Figure 4.1.2 Internal architecture of ATmega328 microcontroller [11] 4.2. Approach Point A, B, C and D were the waypoints and the car had to navigate from point A to point D, passing through point B and C. The car must maintain its couse of travel while navigating through waypoints. An algorithm was developed to constantly compute the car’s direction of travel to true direction of travel. For any deviation in the difference 26 between the above two for more than certain allowable limit would steer the car appropriately compensating for the deviation. Figure 4.2.1 Car mounted with GPS receiver and microcontroller 27 Figure 4.2.1 shows the car used in this project which was mounted with GPS receiver, ATmega328 board, and related circuitry of H-bridge motor controller with separate power supply for the motors and the board. The car was 4-wheel-drive; hence an appropriate differential drive mechanism was used to move the car forward, backward, left and right. The GPS receiver constantly supplied GPS coordinates, course and speed data to the board. The board computed the direction of travel, i.e. weather to go forward, take left or right and provided appropriate signals as input to the H-bridge motor controller. The motor controller in turn drove the motors. The combination of direction of rotation of motors drove the motor forward, backward, left or right. 4.3. Algorithm The project implemented feature of waypoint navigation, using GPS coordinate readings to continuously compute its current direction, distance from the waypoint and true direction. From time to time it corrected itself from any deviation from true direction. Below mentioned are the basic steps implemented for the project. Figure 4.3.1 shows the algorithm. 1. Collect coordinates readings of all waypoints from the GPS receiver 2. Store the coordinates of waypoints into memory (of controller or GPS receiver) 3. Collect coordinates of current position 4. Compute the distance and direction of the next waypoint 5. Start navigating in the set direction 6. Steer car in case it deviates from its course 7. Repeat steps iii to vi until all waypoints are reached 28 In step (i), points A, B, C and D were first decided and multiple readings of latitude and longitude were recorded. The mean of those readings were then used as fixed waypoints for the project. The mean latitude and longitude readings are tabulated in table 3.4.1 and are mentioned here. Point A: 38.5603182128, -121.4210546807, Point B: 38.5603597584,-121.4208304601, Point C: 38.5607323646, -121.4210119247, Point D: 38.5606722938, -121.4211874816. These coordinate points were stored in program structure. Whenever a particular waypoint was achieved, next coordinate point was loaded into main program. The car was kept a little further away from point A. The GPS receiver took 30-40 seconds which is its cold start time to get first readings of the car’s current location. The reason to keep car away from point A was to determine whether the start point A was successfully reached by the car when kept randomly anywhere on the track. Step (iv) is explained in detail in section 4.5. This part of the algorithm determines the distance and direction of next waypoint. Step (v) is driving motors straight or steer left or right based on the computer direction. Whenever the car reached the waypoint in vicinity of 1m, it would consider that particular waypoint had been achieved and move on to next waypoint. As our GPS coordinates showed a precision standard deviation of around 1m, it was decided to consider all points a distance of 1m from waypoint as the waypoint itself. 29 Figure 4.3.1 Algorithm 30 4.4. Computing Distance and Direction of Destination Point This project assumed a spherical model of earth surface. As the trajectory for navigation was limited to few hundreds of meters only, there was a very less chance of a positional error. Equation 4.4.1 gave the distance between two coordinates (ϕ1, λ1) and (ϕ2 ,λ2) in meters. Here R is radius of the earth, which was assumed to be constant as the spherical model of earth surface was under consideration. Also, it is assumed that the positional accuracy of a waypoint achieved would be equal to or less than 1 meter. Hence, if the car reached anywhere in 1 meter radius circle, it was assumed that the waypoint was achieved. d = R.cos −1 (sin φ1.sin φ2 + cos φ1.cos φ2 .cos(λ2 − λ1 )) ……….………..…… (4.4.1) where d = distance between two points R = radius of earth 6371000 m Equation 4.4.2 gave the direction of next waypoint at coordinate (ϕ2 ,λ2) from current position (ϕ1, λ1). This equation gave the direction in range from –π to +π, with 0 degrees set in north direction. 31 θ= atan2 ( sin ( ∆λ ) .cos (φ 2 ) , cos (φ1) .sin (φ 2 ) − sin (φ1) .cos (φ 2 ) .cos ( ∆λ ) ) (4.4.2) x>0 y arctan x y ≥ 0, x < 0 y π + arctan x y y < 0, x < 0 …………….………..……… (4.4.3) atan2( y, x) = −π + arctan x y > 0, x = 0 π 2 π y < 0, x = 0 − 2 undefined y 0,= x 0 = The GPS receiver computed fresh latitude, longitude, course and speed readings at a rate of 1 sec. Hence the values of direction and angle θ were constantly updated. Due to uneven track surface and constantly changing GPS readings, often the car would tend to deviate from its set course towards destination waypoint. The algorithm measured such deviation and if that deviation were to go above or below +45 and -45 respectively, the car was made to take a 45 degree left or a right. The deviation was measured as the numeric difference of angle θ and the course of travel obtained from GPS receiver. The algorithm also updated the deviation with new values of θ and course. Figure 4.4.1 illustrates the scenario. A log file of all latitude, longitude, course and speed readings of GPS receiver was maintained, in addition to logging the distance and deviation computed by the controller. The latitude, longitude readings were extracted from the file and were used to plot a map as described in section 5.1 and 5.2. 32 Figure 4.4.1 Calculating deviation from actual course 33 Chapter 5 EXPERIMENTAL RESULTS 5.1. Plotting the Waypoints on a Map The mean values of latitude and longitudes coordinates shown in table 3.4.1 are plotted on a map, shown in figure 5.1.1. The average distance between point A and point B is 20 m, between point B and point C is 44 m and between point C and point D is 17 m. Considering the precise GPS readings, such distance between the points does not generate any errors in navigation. Figure 5.1.1 Plot of point A, B, C and D on a map 34 5.2. Plotting the Path Traversed By the Car The actual path traversed in comparison to actual path, by the car is shown in figure 5.2.1. Initially, the car takes 2 turns around point A in order to lock on first waypoint. Thereafter it sets off towards point B, its course of direction is not exactly in line with the direction of point B. The algorithm detects as the deviation moves beyond 45 degrees and then turns car to right. This continues till car is on its set course and finally reaches in 1 meter vicinity of point B. At point B again it takes whole 360 degree turn to lock point B and head towards point C. The car sets in direction of point C, intermediately corrected for its course and reaches point C. Again the car takes 360 degree turn and heads towards point D. Point D is the last waypoint, hence the car stops on reaching in vicinity of point D. Thus, the car successfully navigates from point A to point D, via points B and C. 35 Figure 5.2.1 Plot of actual path vs. path traversed by the car 36 Chapter 6 CONCLUSION It was observed that at any time, a current course reading would be largely different from the previous reading. Hence the path traversed by the car is not smooth. There are other factors such as uneven surface that affects the traversed path. Distance between point B and point C is fairly greater compared to other distances. It can be seen that the car navigates smoothly on path from point B to C compared to path between points A to B and C to D. This response can be smoothed by filtering the GPS data received or by getting augmented GPS coordinates. Certain filters like Kalman filters are used in robotics to smooth out the movement of the car. Secondly use of GPS augmentation methods like DPGS or WAAS can also help improve the actual response. 37 BIBLIOGRAPHY [1] A. El-Rabbany, “Introduction to GPS: the Global Positioning System”, 2nd ed., Boston, MA, Artech House, 2006. [2] M. Betke and L. Gurvits, “Mobile Robot Localization Using Landmarks,” 1994 IEEE Proc. on International Conference of Robotics and Automation, Volume 1, Munich, 1994, pp. 135-142. [3] M. Maimone, A. Johnson, Y. Cheng, R. Willson, and L. Matthies, “Autonomous Navigation Results from the Mars Exploration Rover (MER) Mission", 9th International Symposium on Experimental Robotics, Singapore, 2004, pp. 3-13. [4] B. Hofmann-Wellenhof, H. Lichtenegger, and J. Collins, “Global Positioning System, theory and practice”, 5th ed., New York, Sprincer-Verlag, 2001. [5] R. Langley, “Navstar GPS Constellation Status” [Online]. Available: http://gge.unb.ca/Resources/GPSConstellationStatus.txt [6] R. Langley, “The Mathematics of GPS”, GPS World, vol. 2, no. 7, July/August1991. [7] R. Knippers, “Geometric Aspects of Mapping”, educational notes, International Institute for Geo-Information Science and Earth Observation (ITC), Enschede, August 2009. [8] R. Foster, “Basic Geodesy”, National Geospatial-Intelligence Agency, May 2005. [9] SiRF Technology Inc., “NMEA Referene Manual”, rev. 1.3, January 2005. [10] T. Davison, “Navigation Companion”, John Wiley & Sons, 2007. [11] Atmel Corporation, “ATmega48A/48PA/88A/168A/328/328P Summary”, rev. C, August 2010.

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising