MarXbot User Manual Version 1.1 - MOBOTS

MarXbot User Manual Version 1.1 - MOBOTS
MarXbot User Manual
Version 1.1
Philippe R´etornaz, St´ephane Magnenat, Florian Vaussard
Mobots group - LSRO - EPFL
December 12, 2014
Contents
1 The
1.1
1.2
1.3
1.4
1.5
1.6
1.7
MarXbot Robot
Overview . . . . . . . . . . . . . . . . . . . .
Electronics of the Modules . . . . . . . . . . .
Software Description . . . . . . . . . . . . . .
Basics . . . . . . . . . . . . . . . . . . . . . .
1.4.1 Overview . . . . . . . . . . . . . . . .
1.4.2 Power on . . . . . . . . . . . . . . . .
1.4.3 Power off . . . . . . . . . . . . . . . .
Establishing a TCP/IP Connection . . . . . .
1.5.1 Wi-Fi Connection . . . . . . . . . . .
1.5.2 USB Connection . . . . . . . . . . . .
1.5.3 SSH Session . . . . . . . . . . . . . . .
Establishing a Console Session . . . . . . . .
1.6.1 FTDI Module . . . . . . . . . . . . . .
1.6.2 Bluetooth Module . . . . . . . . . . .
Bonus . . . . . . . . . . . . . . . . . . . . . .
1.7.1 Configuration for the Mobots Network
2 Introduction to Aseba Studio
2.1 Getting Aseba . . . . . . . . . . .
2.1.1 Getting the Binaries . . . .
2.1.2 Compiling From Sources . .
2.2 Running Aseba Studio . . . . . .
2.3 Your First Scripts . . . . . . . . .
2.3.1 Light On the RGB LED . .
2.3.2 Your First Local Event . .
2.3.3 Your First Global Event . .
2.4 Advanced Aseba Topics . . . . . .
2.5 Other Aseba Tools . . . . . . . . .
2.5.1 Dashel Targets . . . . . . .
2.5.2 asebacmd . . . . . . . . . .
2.5.3 asebaswitch – asebamedulla
2.5.4 asebadump . . . . . . . . .
2.5.5 asebarec – asebaplay . . . .
2.5.6 asebaeventlogger . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
5
7
7
9
9
9
10
10
10
10
11
11
11
12
13
13
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
14
14
14
14
15
16
16
17
17
19
19
19
19
19
20
20
20
3 Advanced Topics
3.1 Using the Camera . . . . . . . . . . . . .
3.1.1 Selecting the Camera . . . . . . .
3.1.2 Streaming Images . . . . . . . . .
3.1.3 Coding . . . . . . . . . . . . . . .
3.1.4 Advanced . . . . . . . . . . . . . .
3.2 GPIOs and LEDs . . . . . . . . . . . . . .
3.2.1 GPIOs . . . . . . . . . . . . . . . .
3.2.2 LEDs . . . . . . . . . . . . . . . .
3.3 Power Management . . . . . . . . . . . . .
3.4 Accessing the Root Filesystem and Flash
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
21
21
21
21
22
22
22
22
23
23
24
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
24
24
24
24
25
4 Introduction to iMX Programming
4.1 The SDK . . . . . . . . . . . . . .
4.1.1 Installing the SDK . . . . .
4.1.2 Using the Toolchain . . . .
4.2 Remote Debugging . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
26
26
26
27
27
3.5
3.4.1 Micro-SD memory
3.4.2 NOR memory . . .
Updating the Robot . . .
3.5.1 Simple Update . .
3.5.2 The Hard Way . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5 Control Architecture
6 Aseba API
6.1 Calibration . . . . . . . . . . . . . . . . .
6.2 Aseba Application Programmer Interface
6.2.1 Introduction . . . . . . . . . . . .
Visible and Hidden API . . . . . .
Conventions . . . . . . . . . . . . .
Settings . . . . . . . . . . . . . . .
6.2.2 Common API . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Hidden variables . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . .
Hidden Natives Functions . . . . .
Settings . . . . . . . . . . . . . . .
6.2.3 Standard Motor Module . . . . . .
Principle of Operation . . . . . . .
Variables . . . . . . . . . . . . . .
Hidden Variables . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . .
Settings . . . . . . . . . . . . . . .
6.2.4 Treel-left . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . .
6.2.5 Treel-right . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . .
6.2.6 Sensors . . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . .
6.2.7 sensor-turret . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
Natives functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . .
Hidden Variables . . . . . . . . . .
6.2.8 Led-rgb . . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . .
28
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
29
29
29
29
29
29
30
30
30
30
30
30
31
31
31
32
32
33
34
34
34
35
35
36
36
37
37
38
38
38
38
39
39
40
40
40
40
40
40
41
41
41
42
42
42
6.2.9
6.2.10
6.2.11
6.2.12
6.2.13
6.2.14
Natives Functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . . .
Hidden Variables . . . . . . . . . . .
gripper-sensor . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . . .
Hidden Variables . . . . . . . . . . .
Hidden Natives Functions . . . . . .
griper-led . . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . . .
rab2 . . . . . . . . . . . . . . . . . .
Variables . . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . . .
Settings . . . . . . . . . . . . . . . .
Hidden Variables . . . . . . . . . . .
Magnetic-Gripper-Main . . . . . . .
Variables . . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . . .
Magnetic-Gripper-Front . . . . . . .
Variables . . . . . . . . . . . . . . .
Events . . . . . . . . . . . . . . . . .
Natives Functions . . . . . . . . . .
Magnetic-Gripper Motors calibration
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
42
42
42
42
43
43
43
43
43
44
44
44
44
44
44
44
45
46
46
46
46
47
47
47
47
48
48
48
49
49
Foreword
This user manual is devoted to document the marXbot basic use with the bootloaders and aseba virtual
machines already programmed. This is not a documentation about the electronics nor about embedded C
programming on the microcontrollers.
As the marXbot is running an embedded Linux system, most of the tricks in this manual are given for
a Linux host system, apart aseba which is a cross-platform software. When known to work, the Windows
equivalence is given. If you find some working tools, feel free to share with us for future users. Anyway, a
Linux system is strongly encouraged.
You will see the following warnings along the road. Please, always read them and make sure you
understand them. Feel free to ask us any questions you may have, as some operations could easily break
the marXbot.
Mind this point.
This can be harmful for you or the robot. Take great care.
This can destroy the robot. Do this at your own risks.
This can destroy the robot and bring you years of torments, maybe for eternity...
You have been warned.
4
Chapter 1
The MarXbot Robot
This chapter will give you a first crash-course to the marXbot robot, if you are not already familiar with it.
At the end, you should be able to understand the basic architecture, both from the hardware and software
point of view, and establish a first SSH connection with the robot. For a more in-depth description, please
refer to [1–3].
1.1
Overview
An overview of the robot can be found in Fig. 1.1. The marXbot is a general purpose, all-terrain, modular
robot. It has been designed and manufactured by the Mobots group1 , belonging to the EPFL – LSRO2
robotics laboratory.
The robot is modular by design:
1. The base module. It is made up of two wheels (with tracks), the power electronics and a great
number of sensors: 24 IR proximity rangers, IR ground rangers (8 long range + 4 short range), 3D
accelerometer, 3D gyroscope, RFID reader - writer. The power electronics is capable to hot-swap
the battery, allowing the user to exchange the battery without shutting down the power. The main
processor will continue to work, while the less useful parts (motors, sensors,...) will be powered off
to save energy. There should be enough energy to power the processor for 10 seconds.
The hot-swap is achieved using supercapacitors. Thus, a minimum charging
time of 5 minutes should be granted before trying this capability.
2. (Optional) Attachment module. This module allows marXbots to attach to each other in swam
experiments. The gripper can rotate around the robot. This module also provides 12 RGB LEDs.
3. (Optional) Range and bearing module. A marXbot can get the relative position (distance + angle)
of each individual inside a swarm using this module, up to a distance of 5 meters.
4. (Optional) Rotating scanner. This scanner is a low-cost version of common laser range scanners,
using 4 infrared rangers (2 short range + 2 long range). It generates 480 points per second, at a
rotation frequency of 1 to 2 Hz.
5. Main computer with front camera. The main computer is a Freescale i.MX31 processor (ARM
11, 533 MHz, 128 MB RAM), running a lightweight Linux distribution (no graphic desktop). A 3
megapixels camera can be mounted onto the board, in two different orientations (either upward or
frontward).
6. (Optional) Upward camera and beacon LED. This module provides another 3 megapixels camera
and a powerful RGB LED with built-in thermal protection. This camera is mainly used as an
omnidirectional one.
1 http://mobots.epfl.ch
2 http://lsro.epfl.ch
5
Figure 1.1: Overview of the marXbot robot (source: [1])
6
1.2
Electronics of the Modules
The electronics of the modules is briefly described in Fig. 1.2. Each module has up to 3 dsPIC33
microcontrollers, a complete marXbot having a total of 10 microcontrollers. They are interconnected
through a CAN bus. An I2C bus also exists and is used by the i.MX 31 to perform the power management
tasks.
For this reason, the robot can’t be operated without the computer module. This
could be feasible, but solder jumps have to be soldered on the base module in this
case. Ask the Mobots group for this modification.
iMX 31
Wifi
3 mp camera
iMX 31
speaker
microphone
2⨉
dsPIC33
RGB beacon
2⨉
Scanner
rotating platform
distance sensor
rotation motor
dsPIC33
4⨉
induction
induction
Range and bearing
emitters
receivers
dsPIC33
20 ⨉
data (CAN)
energy
dsPIC33
12 ⨉
Gripper module
rotating platform
rotation motor
dsPIC33
gripping motor
traction sensor
dsPIC33
light ring
Base module
IMU / RFID
left treel
right treel
ground sensor below
2⨉
ground sensor below
2⨉
proximity sensor
2⨉
dsPIC33
24 ⨉
ground sensor around
8⨉
Battery pack
protection circuit
LiPoly batteries
Auxiliary power storage for battery hot swap
Figure 1.2: Overview of the electronics (source: [1]).
The CAN architecture is shown is Fig. 1.3 (a). This is a decentralized network, without any router.
The embedded computer is also connected to this CAN network, through a CAN translator (not pictured)
as the iMX31 has no built-in CAN interface. This embedded computer is connected to the outside world
using a TCP/IP socket, operating either over Wi-Fi or USB (usbnet).
1.3
Software Description
The great challenge in the marXbot is to manage the programs distributed between the 10 microcontrollers
and the main CPU. Flashing all those units by hand would be untenable. This is why we take here
advantage of the CAN bus, in order to flash remotely the microcontrollers.
7
Moreover, managing such a network only with low-level programming in C would be hard for the
end-user. This assessment gave birth to aseba, a high-level, event-based scripting language. Aseba
scripting is only available on the microcontrollers. Thus, the program inside each microcontroller is
structured in several layers (Fig. 1.3 (b)):
• A CAN bootloader is responsible for flashing the whole microcontroller, if needed3 .
• The low-level programming is done in C – with some optimizations performed in assembler – allowing
good performances.
• The high-level behavior is done in aseba, leveraging the functions offered by the low-level layer4 .
The aseba script runs inside a small virtual machine. Each script is transferred through the CAN
network and can be permanently written in Flash for persistency purpose.
Figure 1.3: Architecture of the network (a) and microcontrollers (b) (source: [3]).
3 In fact, two microcontrollers don’t have a bootloader, and thus must be flashed using a Microchip ICD2 or ICD3: the
internal translator (UART ↔ CAN, not shown) on the computer module and the rotating dsPIC of the scanner module.
4 Same apply here.
8
1.4
Basics
Here you will learn how to start with the marXbot. Please read carefully what follows.
1.4.1
Overview
Figure 1.4: Top view of the marXbot.
1. Power button + LED
13. ISFC for translator – X14
2. iMX reset button
14. Translator status LEDs
3. Mic 1
15. dsPICs reset button
4. RTC backup battery
16. USB OTG (device)
5. CoreOn LED
17. MarXbot ID tag (Mxx)
6. LED 0–3 (red) + LED 5 (RGB)
18. Wi-Fi dongle
7. ISFC for iMX (UART1) – X11
19. Upward camera
8. MicroSD card
20. Wi-Fi status LED
9. Audio jack
10. Translator reset button
21. Mic 2
11. Selector
22. LED 4 (RGB)
12. SD card slot
23. Front camera
1.4.2
Power on
Insert the battery inside the base module, with the contacts towards the back of the marXbot. The power
LED (1) and core LED (5) will turn on. If nothing happens, press for 1 second on the power button (1).
Shortly after, the heart-beating LED (6) will start to blink. After about 30 seconds, the LEDs of the
translator (14) will also turn on, the marXbot is now ready to be used.
9
1.4.3
Power off
In order to power off the robot, press for 1 second on the power button (1) or execute the poweroff
command on the robot. Wait until the robot shutdowns itself (all LEDs turned off), and then remove the
battery. Please don’t let the battery inside the robot, even if it is powered off.
Always properly shutdown the robot. The file system is of type EXT2 and is thus
not journaled. A sudden power off could corrupt the file system, leaving you with no
other chance than performing a e2fsck on the corrupted SD card. Most of the time,
it will be easier to write a fresh root file system. See Sec. 3.5.2.
Always remove the battery. If you leave it inside, some capacitors may get charged
through pull-up resistors, possibly powering on the microcontrollers and making the
robot to jump down your table, like a Lemming would do. You have been warned.
1.5
Establishing a TCP/IP Connection
The marXbot is a versatile platform, and there are several ways to establish a connection.
1.5.1
Wi-Fi Connection
By default, the marXbot is configured to connect to the WPA-secured “mobots” network, using the
embedded Wi-Fi dongle. This imply the availability of the corresponding access-point. The configuration is summarized in Sec. 1.7.1. The configuration of the robot can of course be changed in
/etc/network/interfaces and /etc/wpa supplicant/.
Supposing that you have the access-point, the LED on the Wi-Fi dongle will become blue when
connected. As the USB driver on the marXbot has some problems, it may happen that the connection is
not established properly, even 1 or 2 minutes after booting. In that case, remove carefully the dongle (18),
insert it again and wait. If you are unlucky, you will have to perform this 2 or 3 times. As a last resort,
reboot the robot using the power button (1).
On your computer, you also need to get associated with this private network, using the information
given in Sec. 1.7.1. The IP address will be allocated by DHCP.
Using this Wi-Fi connection, the marXbot will have the static IP address
IP = 10.0.0.1xx
replacing xx by the number of your marXbot, written on the yellow sticker (17). If your computer is
running zeroconf (avahi for example), you can use the hostname instead of the IP
marxbotxx.local
again replacing xx by the number of your marXbot.
1.5.2
USB Connection
In case the Wi-Fi connection is not available to you, you can create a network over USB using usbnet5 .
To do so, you will need a USB cable (mini-B to std-A). Plug the mini-B into the OTG-HS port (16) of
the marXbot, while connecting the A plug to your computer. If you look at your network interfaces using
ifconfig, you should see a new interface named usb0.
You first need to configure this interface. Your computer should be configured with a address in the
subnet 192.168.0.1/24, for example
sudo ifconfig usb0 192.168.0.1
Now, you should be able to ping the marXbot. They all have the same static IP address
5 On
Windows, you will have to perform a few more steps (not tested): http://docwiki.gumstix.org/index.php/
Windows_XP_usbnet using the .inf provided here http://www.davehylands.com/linux/gumstix/usbnet/linux.inf
10
IP = 192.168.0.202
1.5.3
SSH Session
An SSH (Secure Shell) session will provide you with a remote console, established through a secured
tunnel. Even if this security is not mandatory for common applications, it is a well standardized way. The
following commands are given for a Linux or Mac OS host system. On Windows, you can use PuTTY as
a graphical replacement.
Once the TCP/IP link is established, you can run a remote shell with the command
ssh [email protected] ip
It will ask you for the password, which is empty (just press the return key). If you don’t want to
be asked for the password each time, you can use public key authentication. The public key is already
registered on the robot, you need to register the private key on your computer (to be done at each boot).
The private key mobots-ssh-key is available on our website6 . Run the command
ssh−add ˜/.ssh/mobots−ssh−key
Copying files to the marXbot is easily achieved with the scp command
scp /path/to/file [email protected] ip:[/optional/path]
Don’t forget the “:” when specifying the remote target, even if you omit the remote path (otherwise it
will get copied locally). If you omit the path, it will copy the file in the working directory of the user
(/root/ in this case).
1.6
Establishing a Console Session
If no network is available, for example when experimenting with the Linux kernel or updating the marXbot,
you will have to fall back on a console session. You will need an (emulated) serial port. This is the purpose
of the “ISFC for iMX31 (X11)” connector (7) on the marXbot. Two modules have been developed. The
following instructions are only given for a Linux-like operating system.
1.6.1
FTDI Module
This module (Fig. 1.5) allows you to establish a console session using a USB cable. The white side on the
small male connector gives you the orientation of the pin number 1. It should match the corresponding
mark on the marXbot’s side.
Be sure not to invert the orientation.
When plugged in, you will see a new block device called /dev/ttyUSBx. You can also refer to the
output of the dmesg command for the exact name.
Figure 1.5: FTDI dongle.
We will use minicom as a serial terminal. To enter the configuration screen (only needed the first
time), issue
6 http://mobots.epfl.ch/mx31moboard/mobots-ssh-key
11
sudo minicom −s
and select “Serial port setup”. Enter the setup given in Table 1.1.
Table 1.1: Minicom configuration for USB
A - Serial Device
E - Bps/Par/Bits
F - Hardware Flow Control
G - Software Flow Control
/dev/ttyUSB0
921600 8N1
Yes
No
Before exiting, save the configuration as “ftdi”7 . Launch again minicom, this time using
minicom ftdi
and it should connect.
1.6.2
Bluetooth Module
This module (Fig. 1.6) offers you a wireless serial link. The white side on the small male connector gives
you the orientation of the pin number 1. It should match the corresponding mark on the marXbot’s side.
Be sure not to invert the orientation.
You first have to pair with the Bluetooth device before being able to use it.
hcitool scan
resulting for example in
Scanning ...
00:26:68:0C:C2:6E
10:00:E8:6C:F0:52
00:21:AA:78:4C:FB
08:00:17:2D:09:B2
Nokia 5000d−2
LSRO1 LMX9838 D001
Nokia 6500s−1
e−puck 0012
Copy the 12 digit in front of the device, “LSRO1 LMX9838 D001” in this example. Then
sudo rfcomm bind 0 10:00:E8:6C:F0:52
Don’t forget to replace 10:00:E8:6C:F0:52 with your current device. A new block device called
/dev/rfcomm0 should appear.
Figure 1.6: Bluetooth dongle.
Enter the configuration screen of minicom, exactly as for the FTDI dongle of the previous section, but
using the parameters of Table 1.2.
Save the configuration under the name “rfcomm0” for example. You can then connect using
minicom rfcomm0
7 The
file is usually saved in /etc/minirc.ftdi
12
Table 1.2: Minicom configuration for Bluetooth
A - Serial Device
E - Bps/Par/Bits
F - Hardware Flow Control
G - Software Flow Control
1.7
1.7.1
/dev/rfcomm0
115200 8N1
No
No
Bonus
Configuration for the Mobots Network
The mobots wireless network is using only the 5 GHz band, to avoid the jam on the 2.4 GHz band. So
be sure to use an appropriate access point and Wi-Fi dongle. In our group, we use the Linksys E3000
dual-band router, with the 2.4 GHz band disabled.
Table 1.3: WPA configuration.
SSID
Key type
Key
mobots
WPA-PSK
Ask us
Please, don’t publish the WPA key, as it would compromise the security of the entire
EPFL network. Thank you for your kind comprehension.
Table 1.4: IP configuration.
Router IP
Mask
MarXbot IP range
DHCP range
10.0.0.1
255.255.255.0 (/24)
10.0.0.100-199
10.0.0.200-249
13
Chapter 2
Introduction to Aseba Studio
As you saw in the introduction, the microcontrollers on the marXbot can load and execute aseba scripts. It
is really easy to get started, you only need to install the tools on your computer and start playing. The main
page of the aseba project is hosted on our website at the address http://mobots.epfl.ch/aseba.html.
Aseba is fully cross-platform and has been tested on Windows, Linux and MacOS.
This chapter is not about the aseba language. Please refer to the introductory material available on
the official website, as well as to the API documentation in Chap. 6. An online documentation is also
available in the Help menu of Aseba Studio.
2.1
Getting Aseba
You first have to install aseba, either by installing the binaries or compile the sources. We recommend
you to compile the sources downloaded from SVN, so you will be able to keep the most up-to-date version.
2.1.1
Getting the Binaries
You can download Windows binaries on the aseba website (http://mobots.epfl.ch/aseba.html).
Binaries for Linux (.deb packages) are also available, but are unfortunately outdated. We will try to add
more binaries and keep them up-to-date, but without any warrantee.
2.1.2
Compiling From Sources
The best solution is to build from sources. This is easily achieved thanks to CMake. You will need to
install at least
• cmake
• ccmake
The SVN repository of aseba is hosted on gna1 . Aseba also depends on a number of other libraries:
• Dashel2
• Enki3
• Qt4
• Boost
• Qwt
• SDL
Dependencies other than enki and dashel should be available through the package manager of your
favorite distribution. Take care to install the -devel packages also. For example, on Fedora 14
1 http://gna.org/projects/aseba
2 SVN
3 SVN
repository: http://gna.org/projects/dashel
repository: http://gna.org/projects/enki
14
sudo yum install qt−devel boost−devel qwt−devel SDL−devel
Dashel and enki are to be download on gna using SVN and compiled using CMake (should be easy if
other dependencies are correctly installed)
cmake .
make
sudo make install (optional)
The last step is to compile aseba. The only tricky step is to correctly setup the path to the dependencies.
Most of them should be automatically detected, but you will have to specify by hand the ones for dashel
and enki (if not installed). To configure the paths, run
ccmake .
If the cache is empty, press ’c’ to configure, then ’e’ when errors are displayed. You should then be
able to enter the paths by pressing ’enter’. Typical paths are given in Table 2.1, adapt them to your
configuration. When done, you can rerun the configuration by pressing ’c’ or executing cmake. When the
configuration succeeds, compile as before
cmake .
make
Table 2.1: An example of CMake configuration for aseba
CMAKE INSTALL PREFIX
DASHEL INCLUDE DIR
DASHEL LIBRARY
ENKI INCLUDE DIR
ENKI LIBRARY
ENKI VIEWER LIBRARY
QT QMAKE EXECUTABLE
QWT INCLUDE DIR
QWT LIBRARIES
SDLMAIN LIBRARY
SDL INCLUDE DIR
SDL LIBRARY
2.2
/usr/local
/home/vaussard/svn/dashel
/home/vaussard/svn/dashel/libdashel.a
/home/vaussard/svn/enki
/home/vaussard/svn/enki/enki/libenki.a
/home/vaussard/svn/enki/viewer/libenkiviewer.a
/usr/bin/qmake-qt4
/usr/include/qwt
/usr/lib64/libqwt.so
/usr/lib64
/usr/include/SDL
/usr/lib64/libSDL.so;-lpthread
Running Aseba Studio
The first step is to establish a TCP/IP network with the robot. Follow the instructions given in Sec. 1.5,
either using Wi-Fi or USB (usbnet).
Go into the folder where you have compiled aseba, and run the asebastudio executable
./studio/asebastudio
The connection screen of Fig. 2.1 will show up. Enter the corresponding IP address or hostname, as
described in the introduction. Port should be 33333. If successful (try to ping the robot otherwise), the
Aseba Studio GUI will appear (Fig. 2.2).
Each aseba node (→ each microcontroller) appears in a separate tab. The full documentation for
each node is available in Chap. 6. For a full marXbot, the following nodes are available:
• treel-left: Left wheel and corresponding sensors.
• treel-right: Right wheel and corresponding sensors.
• base-sensors: Range sensors in the base module.
• sensor-turret: Rotating scanner module.
• gripper-sensor : Sensors for the gripper module.
• gripper-led : Gripper’s RGB LEDs and actuation.
15
Figure 2.1: Aseba Studio connection screen.
Figure 2.2: Aseba Studio.
• rab2 : Range and bearing module.
• led-rgb: Power RGB LED.
2.3
Your First Scripts
This section will provide you some basic examples for your first aseba scripts.
2.3.1
Light On the RGB LED
Select the led-rgb tab and enter the following piece of code
call led .rgb(500,0,0)
The compilation into bytecodes is done on-the-fly. Now, you just need to press the Load and Run
buttons on your left. The power RGB LED will turn red! If you prefer green, write instead
call led .rgb(0,500,0)
Load, run and the LED is now green! This powerful LED turns out to heat quite a lot. This is why a
built-in thermal protection has been added. You can measure the LED’s temperature by enabling the
sensor
16
led .temperature.period=100
call led .rgb(0,500,0)
Load, run and press the Refresh button. You will see the measured value in the led.temperature
variable. According to the documentation of Sec. 6.2.8, the unit is 0.1 ◦ C. If you increase the intensity or
enable several colors at the same time, you will see this value increasing each time you press the Refresh
button.
2.3.2
Your First Local Event
Following up the first example, we will slightly increase the complexity in order to make the LED blink at
a frequency of 0.5 Hz, using a local event. Still using the led-rgb tab, copy the Listing 2.1.
# Declare variables
var time = 0
var state = 0
var value
# Set the sampling and timer0 period
led .temperature.period = 100
timers.period [0] = 100
# timer0 event routine
onevent timer0
time = time + 1
if time == 10 then
time = 0
state = ˜state
value = 500∗state
call led .rgb(value ,0,0)
end
Listing 2.1: Second example: blinking LED
A few words of explanation:
• Oneline comments begin with #.
• All the variables are global and should be declared at the beginning.
• Local events are fired by the microcontrollers and cached locally. Each microcontroller has its own
set of local events. They are displayed using the Local Event tab at the lower left corner of the GUI.
• timers.period[0] sets the period (in milliseconds) for the local event fired by Timer0.
• onevent timer0 is the routine executed when the Timer0 event is fired.
2.3.3
Your First Global Event
Global events can be exchanged between the microcontrollers connected to the CAN bus, and also to /
from a computer connected on the same network, like with asebastudio. We will demonstrate this using
the proximity sensors, coupled with the RGB LED to display the detection of an obstacle. The first script,
shown in Listing 2.2, is to be coded in the base-sensors node. In addition, you will have to add a constant
named THRESHOLD set to 1000 (upper right corner). The global events are also to be added in the
Events box, on the right (see Fig. 2.3). The number next to the event is the number of arguments this
event will take. In our case, it is zero.
The code is rather self-explaining, with the help of the API documentation of Chap. 6. The sensors are
refreshed every 100 ms, and the update function tests each sensor for a value greater than THRESHOLD.
If such a case exists, the global event obstacle is emitted.
17
You can already play with this part. Load and execute the script. You will see fired events in the log
window, on the bottom right corner, each time you approach your hand from the base sensors (obstacle
events) and when you go back (no obstacle events).
var i
var nb obstacles
sensors .period = 100
# sensors updated at 10 Hz
onevent sensors.updated
nb obstacles = 0
for i in 0:23 do
if proximity.corrected [ i ] > THRESHOLD then
nb obstacles = nb obstacles + 1
end
end
if nb obstacles > 0 then
emit obstacle
end
when nb obstacles == 0 do
emit no obstacle
end
Listing 2.2: Third example: proximity sensors - base-sensors node
Figure 2.3: Aseba Studio during third example.
You can then catch those events from the led-rgb node, for example to light on the LED. Copy the
script of Listing 2.3 in the corresponding tab.
onevent obstacle
call led .rgb(300,0,0)
onevent no obstacle
call led .rgb (0,0,0)
Listing 2.3: Third example: proximity sensors - rgb-led node
18
2.4
Advanced Aseba Topics
FIXME
2.5
Other Aseba Tools
Aseba is not limited to Aseba Studio. There are several other (command-line) tools, which can be useful
in several situations. We will give a brief summary in this section. Please refer to the output of the
--help switch for detailed help on each command.
2.5.1
Dashel Targets
The following commands use so-called “dashel targets”, which are strings describing the interface to
connect to. They are to be written between quotation marks “ ”. A summary is given in Table 2.2. If
you want to connect to several dashel targets at the same time, you can separate them with with a white
space ( ), like
asebaswitch ”ser:device=/dev/ttyUSB0 tcp:10.0.0.154”
Table 2.2: Dashel targets.
Target type
Serial
TCP / IP
Dashel string
“ser:device=/dev/ttyXX;fc=hard;baud=921600”
“tcp:10.0.0.1XX;port=33333”
The IP can be replaced by localhost, for example when connecting to asebaswitch or asebamedulla.
2.5.2
asebacmd
asebacmd is useful for sending commands over the aseba network. Among the commands, the most
useful are summarized in Table 2.3.
Table 2.3: Useful commands for asebacmd.
Command
presence
usermsg [type] [length]
sb [nodeid]
2.5.3
Usage
Broadcast an identification request.
Send a global event on the network of ID type with length number
of data. For now, the (dumb) data are 0 1 2 ... length-1.
Reset the node with ID nodeid (number).
asebaswitch – asebamedulla
asebaswitch and asebamedulla are almost identical. Both of them are used to connect several aseba
components together, like a control software, a dump program and an aseba network. This will be
explained in greater details in Chap. 5. asebaswitch supports all the dashel targets (serial, TCP).
Moreover, asebamedulla can connect components using D-Bus.
The only limitation is the ID of the nodes4 . Each ID must be unique inside the same network. This
implies that two marXbots can’t connect to the same aseba network, as this would give rise to ID
conflicts.
If you want to leverage D-Bus using your own program and asebamedulla, you can look at the example
Python script, located in the medulla folder. The system bus is used on the marXbot, while the session
bus is used on a remote computer.
For both programs, you can use the -d -v switches in order to display what is going on the network,
which can be quite useful for troubleshooting.
4 The
ID of a node can be found in Aseba Studio by showing the hidden variables.
19
2.5.4
asebadump
asebadump, connected to an asebaswitch, is used to dump all the data going on the aseba network.
Those data are output on the console, using a raw format.
2.5.5
asebarec – asebaplay
asebarec, connected to an asebaswitch, is used to record global events on the aseba network. Those
events are time-stamped and are in a human-readable format.
asebaplay is used to inject recorded events into the aseba network. You can choose to play them
faster than usual.
2.5.6
asebaeventlogger
asebaeventlogger is a Qt interface used to plot a specific event over time. However, this can be also
achieved in Aseba Studio, so it’s of little use.
20
Chapter 3
Advanced Topics
This chapter will cover more advanced usages, like accessing the camera, the LEDs or updating the robot.
3.1
Using the Camera
Up to two Aptina MT9T031 cameras can be mounted on the robot. However, only one acquisition bus
is available on the current processor, thus only one camera can be used at the same time. This can be
partly workarounded, as we will see in Sec. 3.1.4.
The camera appears in Linux as a block device /dev/video0. Operations should be carried on this
node.
3.1.1
Selecting the Camera
The selection between the two cameras is done by an external electronics, activated by one of the GPIO
(General Purpose IO) pin of the i.MX processor. This pin is available in user-space and is called gpio28.
To select the upward camera1 (19)
echo 0 > /sys/class/gpio/gpio28/value
If you prefer to use the frontward camera (23)
echo 1 > /sys/class/gpio/gpio28/value
Never change this configuration while the node is opened. Always close the port
prior to any modifications. In fact, this could mislead the driver, as both the cameras
share the same bus. You must also wait 100 µs between a change on the GPIO and
the opening of the node.
3.1.2
Streaming Images
Streaming images to your computer is the most straightforward operation. First, be sure gstreamer is
installed on your computer. This is easily achieved through the packager manager of your favorite Linux
distribution.
Then, on the robot, launch camera-viewer, which is a custom software installed by default.
./camera−viewer −qws −display VNC:0
Then, on your computer
gst−launch−0.10 tcpclientsrc host=10.0.0.1xx port=11111 ! jpegdec ! autovideosink
Set the IP address to the one of your marXbot (see Sec. 1.5).
You can also set the camera’s parameters (exposure and gain), as camera-viewer offers this facility
by exporting a VNC remote desktop. To access it, use your favorite VNC client (for example KRDC on
KDE desktops) using the IP of your marXbot.
1 This
is the default option.
21
3.1.3
Coding
If you wish to access the camera directly, the best way is to process the images on the iMX. Any program
can get access to the camera, using the Video4Linux2 API. The documentation of this API is unfortunately
quite messy. The camera-viewer application is a good start, ask us for the source.
3.1.4
Advanced
As said earlier, two connectors are available for connecting up to two cameras at the same time. However,
there is only one acquisition bus, shared between the two cameras. Thus, we strongly recommended to
use only one camera at the same time.
With some tricks, it is however possible to access both alternatively, but this is strongly discouraged
and won’t be supported.
Now stop joking, it is really really really discouraged...
To do so, you will need to write your own acquisition program (previous section), in order to synchronize
the operations. Follow the following steps.
1. Select the first camera.
2. Wait at least 100 µs.
3. Open and configure the node.
4. Select the other camera.
5. Wait at least 100 µs.
6. Close.
7. Open and configure the node with the same configuration.
8. Now you should be able to alternate the acquisition on both cameras, by changing the GPIO and
waiting between each acquisition. Don’t change the GPIO in the middle of an acquisition, really...
3.2
GPIOs and LEDs
On the marXbot, GPIOs (General Purpose Input Output) and LEDs are exported by the kernel. This is
a convenient way to interact with the hardware level.
3.2.1
GPIOs
GPIOs can be accessed on the marXbot through the /sys/class/gpio/* pseudo-nodes. Table 3.1 shows
available GPIOs and their usage.
An output GPIO can be put high with
echo 1 > /sys/class/gpio/??/value
and put low with
echo 0 > /sys/class/gpio/??/value
For input GPIOs, you can read the value using
cat /sys/class/gpio/??/value
22
Table 3.1: Available GPIOs
Name
gpio21
gpio22
gpio28
gpio40
gpio41
gpio42
gpio43
gpio87
3.2.2
Direction
Output
Output
Output
Input
Input
Input
Input
Input
Function
Translator reset
dsPICs reset
Select the camera
Selector bit 0
Selector bit 1
Selector bit 2
Selector bit 3
Battery detection
LEDs
LEDs can be accessed on the marXbot through the /sys/class/leds/* pseudo-nodes. Table 3.2 shows
available LEDs and their location. For each node, the actual brightness is retrieved with
cat /sys/class/leds/??/brightness
and the maximum brightness with
cat /sys/class/leds/??/max brightness
while the brightness is set with
echo brightness > /sys/class/leds/??/brightness
Replace brightness with the brightness you desire.
Triggers can also be set. This way allows you to automatically turn on / off the LEDs, based on some
system events. A list of available triggers for each LED is obtained with
cat /sys/class/leds/??/trigger
and a trigger can be set (for example heartbeat)
echo heartbeat > /sys/class/leds/??/trigger
Table 3.2: Available LEDs
Name
ar9170-phy2::assoc
ar9170-phy2::tx
coreboard-led-0:red:running
coreboard-led-1:red
coreboard-led-2:red
coreboard-led-3:red
coreboard-led-4:red
coreboard-led-4:green
coreboard-led-4:blue
coreboard-led-5:red
coreboard-led-5:green
coreboard-led-5:blue
3.3
Location
Wi-Fi dongle (color?)
Wi-Fi dongle (color?)
iMX processor board (heart-beat)
iMX processor board
iMX processor board
iMX processor board
Front camera (RGB → R)
Front camera (RGB → G)
Front camera (RGB → B)
iMX processor board (RGB → R)
iMX processor board (RGB → G)
iMX processor board (RGB → B)
Power Management
You can read the battery voltage on the marXbot using the command:
qdbus −−system ch.epfl.mobots.power /battery ch.epfl.mobots.power.battery.GetVoltage
23
3.4
Accessing the Root Filesystem and Flash
The files on the marXbot are stored either on the micro-SD card (root filesystem), or in the embedded
NOR memory (Linux kernel). The way to access those files is highly depend on the storage media.
3.4.1
Micro-SD memory
The filesystem on the micro-SD card can be easily accessed either through the network (Sec. 1.5), or
directly by removing the SD card (halt the robot before proceeding) and combined with a card-reader.
3.4.2
NOR memory
The embedded NOR memory has the following layout.
Table 3.3: Typical NOR layout (may change).
Node
mtdblock0
mtdblock1
mtdblock2
mtdblock3
Name
RedBoot
kernel
FIS directory
RedBoot config
Size [kB]
256
2048
124
4
Read-only
No
No
Yes
Yes
Only the Linux kernel and the bootloader partition is writable. The other partitions are forced to
read-only.
Please, don’t mess the bootloader partition.
3.5
Updating the Robot
There are tons of ways to screw up the update. So don’t do it unless needed. If you
have any doubt or any potential misunderstanding, please, ask before performing the
update.
3.5.1
Simple Update
If the robot has access to the Internet using the Wi-Fi, this simple update process is enough. It will
update everything, including
• The Linux distribution
• The Linux kernel on the NOR
• The firmware of the microcontrollers2
Using a remote console (SSH or serial), issue the commands
opkg update
opkg upgrade
If you want to update the microcontrollers’ firmware, first make sure the battery is not empty. Then
update−robot
2 Excepted
the 2 microcontrollers without a bootloader, see Sec. 1.3
24
DO NOT, for *ANY REASON* kill this task (or do ctrl-c).
3.5.2
The Hard Way
This way may be your only chance, if the distribution on the robot is really old (without the Wi-Fi
support) or if your SD card is corrupted.
Do this only, only, only if needed. Ask us if you are unsure. This can leave the robot
unusable if you screw up the kernel and we will have to load a new kernel by hand.
Updating the SD Card First download the script located here http://mobots.epfl.ch/mx31moboard/
create_sd.sh. Change the permissions to allow its execution3 . This script assume that your SD card is
already properly formatted. Insert the SD card into your computer. The following commands, executed
from the directory where you have downloaded create sd.sh, will help you to download the latest
available root filesystem. The tarball must be renamed to img.tar.bz2 for the script to work.
wget http://mobots.epfl.ch/mx31moboard/latest rootfs
wget −i latest rootfs
FILE=”ls Angstrom∗”
mv $FILE img.tar.bz2
install −d mnt
sudo ./create sd footbot MARXBOT NUMBER /dev/partition of the sd card
replacing MARXBOT NUMBER by the two-digits number of the marXbot (yellow sticker) and /dev/partition of the sd card by the device of the SD card (usually something like [FIXME]). This will format the
SD card, copy the image and then do some configuration stuff.
A few important points:
• Keep the marXbot number correct. This is important to keep track of each robot and to assign the
correct IP address.
• You must use the first partition of the SD card like [FIXME] (the SD cards on the robots are already
correctly partitioned for this), not the whole device ([FIXME]).
Updating the Kernel To update the kernel, boot it with
• The new SD card
• A charged battery
• The Bluetooth module on the X11 connector (see Sec. 1.6)
Plug the battery in the robot, connect to the Bluetooth and wait until the robot displays the login
prompt (it can take some time for the first boot).
Log in as root (empty password) and do
cat /boot/zImage > /dev/mtdblock1
When the command is done (it will take some time, writing inside the NOR memory is slow), reboot
reboot
It should now be able to connect to the Wi-Fi or usbnet. If it does not connect to the Wi-Fi, reboot
until it succeeds (wait at least 1 minute after each boot to wait for the connection).
Now, you can proceed with the regular update process as explained in Sec. 3.5.1.
3 chmod
+x create sd.sh
25
Chapter 4
Introduction to iMX Programming
Programs can be developed for the embedded computer. This allows you for example to [FIXME]. However,
this embedded processor is based on an ARM architecture, implying the use of a cross-compiler to generate
the binaries. This part can only be performed on a Linux host system, as the SDK is not available for
other operating systems.
4.1
The SDK
The SDK includes several pieces of software, so you can develop, investigate and debug programs running
on the ARM processor. This includes
• The cross-gcc toolsuite (arm-angstrom-linux-gnueabi-gcc and so on...)
• The CMake and Qt tools (cmake, qmake,...)
• Cross-compiled standard libraries (libc, libstdc++,...)
• Other cross-compiled libraries you will need to link most of the programs we provide ()
4.1.1
Installing the SDK
To find out the latest version of the toolchain, first download the manifest from our website
wget http://mobots.epfl.ch/mx31moboard/sdk/latest sdk
and use the link provided in the file to download the SDK corresponding to the architecture of your
development computer. For example, with a 64 bits computer
wget http://mobots.epfl.ch/mx31moboard/sdk/angstrom−2010.4−test−20100608−x86 64−linux−
armv6−linux−gnueabi−toolchain−mobots.tar.bz2
If you want to download the SDKs for all architectures
wget −i latest sdk
Once done, you can install it on your development machine. To install it, just decompress the archive
at the root of your computer
sudo tar −xjf toolchain . tar .bz2 −C /
It will install its files in /usr/local/angstrom/arm. You then need to source the environment-setup
script at startup, for example from your .bashrc, in order to set a number of environment variables for
the SDK to work. To achieve this, just execute once
echo source /usr/local/angstrom/arm/environment−setup >> ˜/.bashrc
26
4.1.2
Using the Toolchain
You can then invoke the cross-GCC using the arm-angstrom-linux-gnueabi- prefix. In order to make your
life easier, a CMake file is freely available at http://mobots.epfl.ch/mx31moboard/sdk/moboard.cmake.
It will set the necessary variables for a smooth compilation, like this:
cmake −DCMAKE TOOLCHAIN FILE=˜/moboard.cmake .
Once compiled, simply copy the binary to the target and run it.
4.2
Remote Debugging
The toolchain includes gdb (called arm-angstrom-linux-gnueabi-gdb) and it can be used to debug
programs running on the marXbot from your computer. To achieve this, we use remote cross-target
debugging with gdbserver.
First you need to compile your program with the debug option (-g with gcc) so that it contains the
debug symbols. With CMake, you need to add the -DCMAKE BUILD TYPE=Debug option. If we go
back to the above command, it becomes:
cmake −DCMAKE BUILD TYPE=Debug −DCMAKE TOOLCHAIN FILE=˜/moboard.cmake .
On the marXbot, run your program with gdbserver and the port the server is going to listen to for
the TCP connection:
gdbserver :5000 yourprogram
On your computer, just start your program with the toolchain’s debugger:
arm−angstrom−linux−gnueabi−gdb
If gdb complains about not being able to load libexpat.so.0, just create a link /usr/lib/libexpat.so.0 to
/usr/lib/libexpat.so.
First, you need to tell gdb where the libraries are on your computer. Here is the default SDK install
path:
(gdb) set solib −search−path /usr/local/angstrom/arm/arm−angstrom−linux−gnueabi
Then tell gdb to read the symbols from your debug enabled program:
(gdb) symbol−file yourpgram
Then you just need to issue the target command in gdb on your computer so that it connects to the
gdbserver on the robot:
(gdb) target remote marXbot ip:5000
From this point, you can use gdb as if you were debugging on your computer. Please note that not all
libraries have the debug symbols enabled.
27
Chapter 5
Control Architecture
28
Chapter 6
Aseba API
6.1
Calibration
The marXbot should already be calibrated.
Any new marXbot must be calibrated prior to first use. To do so, use the relevant python script. This
script uses asebamedulla and thus depends on D-Bus to run. A recent Linux distribution should run
D-Bus by default. The procedure to follow will be displayed by the python script. Make sure to do the
calibration in an environment without external infrared sources (such as windows exposed to sun, lamps,
etc).
6.2
Aseba Application Programmer Interface
6.2.1
Introduction
All the marXbot modules have common interfaces available on any module and some specific to each
modules. The specific interfaces will be discussed in each module’s subsection.
Visible and Hidden API
The API can be subdivided in two major parts:
Visible Part
This subset of the API is safe to use. Some range checking is done on the variables and natives
functions. It should not be possible to crash the microcontroller when using it.
Hidden Part
This subset is unsafe to use blindly. You must double-check what you are doing
when using such variables or natives functions. You can crash the microcontroller,
burn it or damage the whole robot.
Such variables or natives functions are hidden by default in aseba Studio1 . These begin with a “ ”
or have “ .” in their names and will be displayed in gray in the rest of this manual.
Conventions
Some natives functions and variables uses angles. In aseba, the angles are 16 bit integers where −32768
is equal to −π, 0 is a null angle, and 32767 is almost equal to π.
In the documentation of natives functions, variables in capitals letters denote vectors, while variables
in minuscules denote scalars. Variables in capitals letters with indices, such as Ai means for all i in the
range of indices of A.
1 You
can change this by checking the Settings→Show hidden variables and functions option
29
Settings
A hidden functionality of aseba is the settings subsystem. In almost every microcontrollers some
calibration constants must be stored across powerdown. These settings should not be modified without
special care. They are stored inside the flash of the microcontroller and thus only a limited number of
write cycle is allowed.
6.2.2
Common API
Variables
event.source
The aseba node identifier of the incoming event. Only valid for external events. If you write it, you
loose the identifier of the last received event.
event.args
The arguments of the incoming event. Only valid for external events. If you write it, you loose the
arguments of last received event.
Hidden variables
id
The aseba node identifier of the current virtual machine. Unsafe to write.
fwversion
The firmware’s version of the microcontroller. Unsafe to write.
Events
There is no common events.
Natives Functions
math.copy(A,B)
Copy the vector B in the A vector, element by element: Ai = Bi .
math.fill(A,c)
Fill each element of the A vector by the constant c: Ai = c.
math.addscalar(A, B, c)
Compute Ai = Bi + c where c is a scalar.
math.add(A, B, C)
Compute Ai = Bi + Ci where A, B and C are three vectors of the same size.
math.sub(A, B, C)
Compute Ai = Bi − Ci where A, B and C are three vectors of the same size.
math.mul(A, B, C)
Compute Ai = Bi · Ci where A, B and C are three vectors of the same size.
Warning: This is not a dot product.
math.div(A, B, C)
Compute Ai = Bi /Ci where A, B and C are three vectors of the same size.
Note: An exception will be trigged if a division by zero occurs.
math.min(A, B, C)
Write the minimum of each element of B and C in A where A, B and C are three vectors of the
same size: Ai = min(Bi , Ci ).
math.max(A, B, C)
Write the maximum of each element of B and C in A where A, B and C are three vectors of the
same size: Ai = max(Bi , Ci ).
30
math.dot(r, A, B, n)
Compute the dot product between two vectors of the same size A and B:
P
(Ai · Bi )
r= i n
2
math.stat(V, min, max, mean)
Compute the maximum, the mean and the minimum value of vector V .
math.muldiv(A, B, C, D)
Compute multiplication-division using internal 32 bit precision: Ai =
Bi ·Ci
Di .
Note: An exception will be trigged if a division by zero occurs.
math.atan2(A, Y, X)
Yi
Compute Ai = arctan X
using the signs of Yi and Xi to determine the quadrant of the output,
i
where A, Y and X are three vectors of the same size.
Note: Xi = 0 and Yi = 0 will produce Ai = 0.
math.sin(A, B)
Compute Ai = sin(Bi ) where A and B are two vectors of the same size.
math.cos(A, B)
Compute Ai = cos(Bi ) where A and B are two vectors of the same size.
math.rot2(A, B, angle)
Rotate the vector B by angle, write result to A.
Note: A and B must be two vectors of size 2.
math.sqrt(A, B) √
Compute Ai = Bi where A and B are two vectors of the same size.
math.nzseq(a, B, m)
Write to a the middle index of the largest sequence of non-zero elements from B, −1 if not found or
if the sequence is smaller than m.
Hidden Natives Functions
system.reboot
Hard-reset the microcontroller immediately. The microcontroller will execute the bootloader
immediately after the reset.
system.settings.read(a, v)
Read the setting number a and put its value in v.
system.settings.write(a, v)
Write the setting number a with value v.
system.settings.flash
Flash the settings inside the microcontroller’s flash.
Warning: This is really dangerous as the flash wears every time you write to it.
Settings
There is no common settings.
6.2.3
Standard Motor Module
This is the interface used by most motors on the marXbot. It comprises the set of variables, events and
natives functions which manage the PID controllers of the motors. Each motor has his own name and
there can be multiple motors on the same aseba node. Thus the variables, natives functions and event
names are prefixed by the motor’s name. We will use the following name in this section: “M”.
31
Figure 6.1: Standard Motor Module PID architecture.
Principle of Operation
The controller of the motor module is a triple PID controller nesting current, speed and position subcontrollers. Thus there are one PID for setting the current to the motor, one to set the speed, and one to
set the position, as shown in Figure 6.1.
This architecture is flexible as it allows to set limits on maximum current or maximum speed. One
can disable the position controller and directly feed the input of the speed controller to directly control
the motor speed. One can apply the same override to the current controller. The following pseudocode
shows the generic code of the PID controller. The PD or PI controllers are architecturally similar with
the corresponding term set to zero.
i(t) = i(t − 1) + e(t)
output =
(6.1)
e(t) · Kp + i(t) · Ki + Kd · (e(t) − e(t − 1))
scaler
(6.2)
Moreover the speed and position time constants are not in the same order of magnitude than the
current ones. Thus the current controller must run at a higher frequency than the speed and position
controllers.
The current controller filters the current through an IIR filter with the same time constant as the
motor thermal time constant. This ensures that the current imposed to the motor is never higher than its
nominal current.
Variables
M.pid.current max
The absolute value of the maximum current in the motor in milliampere. If higher than the nominal
current the controller will internally automatically reduce it to avoid overheating. This allows to set
the maximum torque of the motor.
M.pid.speed max
The maximum speed of the motor. The unit depends on implementation.
Warning: The maximum speed is not enforced if the controller is set in current mode.
M.pid.enable
The selection of the controller:
0
The controller is disabled, the motor is short-circuited.
1
Only the current controller is enabled. Use M.pid.target current to set the target current.
2
The speed control is enabled (thus the speed and current controllers are enabled). Use
M.pid.target speed to set the target speed.
3
The position control is enabled (thus the speed, current and position controllers are enables).
Use M.pid.target position to set the target position.
M.current
The actual current in the motor, in milliampere. Only updated if the PID controller is enabled.
M.speed
The actual speed of the motor. The unit depends on implementation. Only updated if the PID
controller is enabled.
32
M.position
The actual position of the motor. The unit depends on implementation. Only updated if the PID
controller is enabled.
M.pid.target current
The target current of the current controller, in milliampere.
M.pid.target speed
The target speed of the speed controller. The unit depends on implementation.
M.pid.target position
The target position of the speed controller. The unit depends on implementation.
Hidden Variables
You should not need to read nor to write these varibles in the basic Standard Motor Module use case.
M.pid. period
The period, in millisecond, of the current controller. If positive and M.pid.enable is 1, 2 or 3 the
PID is executing. If negative the PID is not executing and M.pwm is directly applied without any
safeguard on the motor. The M.pid.timer event is generated at the specified period (absolute
value). The maximum period is 400 ms.
M.pid. kp i
The Kp term of the current controller, must be positive.
M.pid. ki i
The Ki term of the current controller, must be positive.
M.pid. scaler i
The scaler term of the current controller, must be ≥ 0.
M.pwm
The PWM applied to the motor.
M.pid. prescaler s
The period multiplier of the speed and position controller. Must be > 0.
M.pid. current speed pulse
The actual speed of the motor in raw units.
M.pid. target speed pulse
The target speed of the motor in raw units.
M.pid. kp s
The Kp term of the speed controller, must be positive.
M.pid. ki s
The Ki term of the speed controller, must be positive.
M.pid. kd s
The Kd term of the speed controller, must be positive.
M.pid. scaler s
The scaler term of the speed controller, must be ≥ 0.
M.pid. speed max pulse
The maximum speed in raw units. This value is automatically overwritten each time M.pid.speed max
is changed.
M.pid. kp p
The Kp term of the position controller, must be positive.
M.pid. kd p
The Kd term of the position controller, must be positive.
33
M.pid. scaler p
The scaler term of the position controller, must be ≥ 0.
M.pid. target position pulse
The target position in raw units.
M.pid. nominal current
The nominal current of the motor, in milliampere.
M.pid. time cst
The thermal time constant of the motor, in millisecond.
M.pid. override
Set to one allow the motor to exceed the software-defined bounds.
M.enc. pulse
The motor’s position in raw units.
M. raw current
The motor’s current in raw units.
Events
M.pid.timer
This event is triggered at the end of execution of every speed controller.
Note: Even if the motor controller is set to current mode, the event will be triggered at the rate of
the speed controller.
M.overcurrent
This event is triggered when the mean current of the motor is higher than its nominal current. This
implies that the maximum current of the motor is reduced to the nominal current of the motor.
M.overcurrent.cleared
This event is triggered after M.overcurrent when the mean current of the motor is back lower than
90% of the nominal current.
Natives Functions
There is no native function for the Standard Motor Module.
Settings
You should not alter the settings, unless you know what you are doing. All the position are relative to the
start of the Standard Motor Module section.
0
Kp of the current controller.
1
Ki of the current controller.
2
Scaler of the current controller.
3
Prescaler of the speed and position period.
4
Kp of the speed controller.
5
Kd of the speed controller.
6
Ki of the speed controller.
7
Scaler of the speed controller.
8
Maximum current allowed in the motor, in milliampere.
9
Nominal current of the motor, in milliampere.
10
Motor’s heating time constant, in millisecond.
34
11
Maximum speed allowed, unit is implementation dependant.
12
Kp of the position controller.
13
Kd of the position controller.
14
Scaler of the position controller.
15
Period of the current controller, in millisecond.
16
Raw offset of the current.
17
Software maximum bound on position.
18
Software minimum bound on position. If the maximum and minimum positions are equals, the
software end stops are disabled.
6.2.4
Treel-left
This aseba node manage the following sensors and actuators:
• Left treel2 .
• Two left-side, close range, infrared ground sensors.
• Three axis accelerometer.
• Three axis gyroscope.
• 15.56Mhz ISO 15693 RFID host.
Due to some internal limitation the two infrared ground sensors, the gyroscope and the accelerometer
are bound to the same sampling speed.
This module embedd a standard motor module named motor. The speed and distance unit is the
motor “pulse”. For a more convenient unit, a special virtual encoder is available which unit is the thenth
milimeter. Please refer to section 6.2.3 for detailed documentation of this module.
Variables
motor.enc.period
The period in ms of the virtual encoder. Maximum is 400ms. 0 mean disabled.
motor.enc.delta pulse
The number of motor pulse (raw unit) since the last virtual encoder event.
motor.enc.delta dist
The traveled distance in
1
10
mm since the last virtual encoder event.
sensors.period
The update period of the ground and imu sensors. Maximum is 400ms, minimum is 10ms. 0 mean
disabled.
ground.corrected
The calibration-corrected value of the two ground sensors. The calibration should produce a value
of 1000 when the robot is on a white paper sheet.
ground.ambiant
The ambiant infrared value. This value is read just before switching on the infrared LED of the
sensor.
ground.reflected
The reflected infrared value. This value is read 300us after the infrared LED of the sensor has been
switched on.
2 Wheel
and track
35
ground.delta
This is strictly equal to ground.reflected minus ground.ambiant.
imu.enable
0 mean accelerometer and gyroscope are powered off. Write 1 to power them on.
imu.acc.corrected
The three accelerometer axis in X, Y and Z order without offset. The unit is ≈ 993 code/g. The
range is ±1.5g
imu.acc.raw
The three accelerometer axis in X, Y and Z order without offset cancelation. The unit is the same
as imu.acc.corrected.
imu.gyro
The three gyroscrope axis. Thoses value have an offset which change with temperature. The drift
◦
become noticeable after about 2-5 minutes. The unit is ≈ 31.2 code/ s . The manufacturer of the
gyroscope give a -50%, +100% tolerance on the scale factor.
rfid.enable
0 mean the rfid chip is powered down. Write 1 to power it up. You must power up the rfid chip
before using any of the rfid native functions.
timers.period
The period of four general purpose timers. 0 mean disabled. Maximum period is 400ms. The timerX
event expire each time the period timeout.
interaxis
The distance between the two wheel of the robot in
1
10
mm.
Events
motor.enc.timer
This event is trigged each time the virtual encoder timer expire. See the variables motor.enc.* for
more informations.
sensors.updated
This event is trigged each time new sensors values are available. See the variables sensors.period,
ground.*, imu.gyro and imu.acc.*.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
rfid.scan done
This event is trigged when the scanning operation of the rfid chip is over. You can check the result
of scan with the rfid.count() native function. See also the rfid.scan() native function.
rfid.read done
This event is trigged when the read operation on a rfid tag is done. The result of the scan is now
available in the argument data passed to the rfid.read() native function.
rfid.write done
This event is trigged when the write operation on the rfid tag is done. The status (success/failure)
is available in the failed argument passed to the rfid.write() native function.
Natives functions
rfid.scan
Start a RFID scan. This will initiate a standard enumeration of all the RFID tags in close contact
with the antenna. A maximum of five tag will be discovered, this should not be a problem as the
antenna can power a maximum of 3 tags. This native function will trigger the rfid.scan done
event when the enumeration process is over.
36
rfid.count(n)
Get the number of RFID tags found by the last enumeration.
rfid.result(n, id, rssi, block, count)
Get the ID, received signal strength indication, block size and block count of the tag number N.
n Tag number. Must be < to the number returned by rfid.count().
id The 64bits unique identifier of the tag.
rssi Received signal strength indication. Minimum value: 0, maximum value: 31.
block The tag block size in bytes.
count The tag block count.
rfid.read(i, b, d, failed)
Initiate a read operation of the EEPROM block b of the RFID tag with unique identifier i. When
finished the rfid.read done event is generated. If failed is equal to 0, the data read are placed in
the d If failed is nonzero the read has failed, thus d is noninitialised.
Note: The variables failed and d must not be touched between the call to rfid.read() and the
rfid.read done event.
rfid.write(i, b, d, failed)
Initiate a write operation to the EEPROM block b of the RFID tag with unique identifier i. When
finished the rfid.write done event is generated. If failed is equal to 0, the data has been succefully
written. If failed is nonzero the write operation has failed.
Note: The variables failed and d must not be touched between the call to rfid.write() and the
rfid.write done event.
Settings
The settings number 0 to 18 are used by the standard motor module.
19, 20
Ground sensors calibration offset.
21, 22
Ground sensors calibration gain.
23
Distance calibration value. Number of motor pulse per straight 50cm displacement of the robot.
24
Rotation calibration value. Number of motor pulse per full trun over itself of the robot.
25 . . . 27
Accelerometer offsets.
6.2.5
Treel-right
This aseba node manage the following sensors and actuators:
• right treel.
• Two right-side, close range, infrared ground sensors.
• Supercapacitor charge.
Due to some internal limitation the two infrared ground sensors and the capacitor charge voltage are
bound to the same sampling speed.
This module embedd a standard motor module named motor. The speed and distance unit is the
motor “pulse”. For a more convenient unit, a special virtual encoder is available which unit is the thenth
milimeter. Please refer to section 6.2.3 for detailed documentation of this module.
37
Variables
motor.enc.period
The period in ms of the virtual encoder. Maximum is 400ms. 0 mean disabled.
motor.enc.delta pulse
The number of motor pulse (raw unit) since the last virtual encoder event.
motor.enc.delta dist
The traveled distance in
1
10
mm since the last virtual encoder event.
ground.period
The update period of the ground sensors and the capacitor charge. Maximum is 400ms, minimum is
10ms. 0 mean disabled.
ground.corrected
The calibration-corrected value of the two ground sensors. The calibration should produce a value
of 1000 when the robot is on a white paper sheet.
ground.ambiant
The ambiant infrared value. This value is read just before switching on the infrared LED of the
sensor.
ground.reflected
The reflected infrared value. This value is read 300us after the infrared LED of the sensor has been
switched on.
ground.delta
This is strictly equal to ground.reflected minus ground.ambiant.
battery.capacitor
The charge voltage of the supercapacitors. The unit is the mV.
timers.period
The period of six general purpose timers. 0 mean disabled. Maximum period is 400ms. The timerX
event expire each time the period timeout.
interaxis
The distance between the two wheel of the robot in
1
10
mm.
Events
motor.enc.timer
This event is trigged each time the virtual encoder timer expire. See the variables motor.enc.* for
more informations.
ground sensors.updated
This event is trigged each time new sensors values are available. See the variables ground.* and
battery.capacitor.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives functions
There is no native functions specific to this virtual machine.
Settings
The settings number 0 to 18 are used by the standard motor module.
19, 20
Ground sensors calibration offset.
38
21, 22
Ground sensors calibration gain.
23
Distance calibration value. Number of motor pulse per straight 50cm displacement of the robot.
24
Rotation calibration value. Number of motor pulse per full trun over itself of the robot.
6.2.6
Sensors
This aseba node manage the following sensors:
• 24 horizontal infrared proximity bumpers.
• 8 vertical infrared sensors.
As the differents proximity sensors must be synchronized to ensure they don’t interfer with each others,
they are bound to the same sampling rate.
Variables
proximity.ambiant
The raw ambiant infrared value of the horizontal sensors. The sensor number is counted clockwise
starting from the front of the robot. This value is read just before switching on the infrared LED of
the sensor.
proximity.reflected
The raw reflected infrared value of the horizontal sensors.
proximity.delta
This is equal to proximity.reflected minus proximity.ambiant.
proximity.corrected
This is the processed sensor value. All the sensors should have a near zero offset when no object is
present and a somewhat identical gain.
ground.ambiant
The raw ambiant infrared value of the vertical sensors.
ground.reflected
The raw reflected infrared value of the vertical sensors.
ground.delta
This is strictly equal to ground.reflected minus ground.ambiant.
ground.corrected
This is the processed sensor value. All the sensors should have a near zero offset when no object is
present and a somewhat identical gain.
sensors.period
The sampling period of the sensors in ms. The maximum value is 400ms, the minimum is 5ms. 0
mean the sensors are disabled.
sensors.bitfield
This is a two word (32bits) bitfield where a 1 indicate the sensor is switched on and 0 mean the
sensor will not be switched on during the scanning process. The bits number 0 to 23 are used for
the horizontal sensors and the bits 24 to 31 are used for the vertical sensors.
timers.period
The period of six general purpose timers. 0 mean disabled. Maximum period is 400ms. The timerX
event expire each time the period timeout.
39
Events
sensors.updated
This event is trigged each time new sensors values are available.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives functions
There is no native functions specific to this virtual machine.
Settings
0 . . . 23
Horizontal sensors calibration offset.
24 . . . 47
Horizontal sensors calibration gain.
48
Horizontal sensors clamp value.
49 . . . 56
Vertical sensors calibration offset.
57 . . . 64
Vertical sensors calibration gain.
6.2.7
sensor-turret
This aseba node manage the following sensors and actuators:
• Two long range infrared sensors (20cm to 1.5m).
• Two short range infrared sensors (0cm to 30cm).
• Rotating motor.
The sampling rate of the sensors is fixed to 60Hz.
Variables
sharp.value
The raw value from the sensors.
sharp.dist
The computed distance in mm.
sharp.angle
The position of the sensors in degree.
voltage
The voltage of the secondary controller.
Events
sharp.updated
New sensors values are available.
undervoltage
The secondary controller experienced an undervoltage condition. The sensors value can be degraded.
sharp.disconnect
The secondary controller had a communication or power failure, thus disconnected the sensors. You
can restart it with sharp.start().
40
Natives functions
sharp.start
Start to power the sensors. You can call the sharp.set speed, sharp.set position or sharp.send conf
before powering up the sensor.
sharp.stop
Power down the sensors. Call this when you don’t need the sensors anymore as it consume quite a
large amount of power (about 2W).
sharp.set speed(s)
Set the sensors’ rotating speed s in RPM.
sharp.set position(p)
Set the sensors’ position p in degree.
sharp.send conf
Force the configuration transfert to the sensors. The natives functions sharp.set speed and
sharp.set position already schedule a such transfert.
Settings
0 . . . 10
X calibration value for sensor 0
11 . . . 21
Y calibration value for sensor 0
22 . . . 32
X calibration value for sensor 1
33 . . . 43
Y calibration value for sensor 1
44 . . . 54
X calibration value for sensor 2
55 . . . 65
Y calibration value for sensor 2
66 . . . 76
X calibration value for sensor 3
77 . . . 87
Y calibration value for sensor 3
Hidden Variables
motor. command
The raw value to send to the motor. If position control is wanted the command should be computed
as:
−position − 1
. The position must be given in raw unit. If speed control is required the command should be
computed as
speed + 16384
. The speed must be given in raw unit.
sharp. led
The led used for the scan for each sensor. Each item in the array is a 8 bits number with the first
nibble correspondig to the first led to use and the second to the second led. The led are numeroted
from 1 to 5. Use 0 to switch off one led. The default value is 3.
sharp. raw position
The position of the sensor in raw unit.
41
6.2.8
Led-rgb
This aseba node manage the following sensors and actuators:
• One RGB 3W beacon LED.
• One LED’s temperature sensor.
The temperature sensor is thermally coupled to the led. If the temperature is above 80◦ C, the led will
switch off automatically and emit the led.overheat event. If will return to normal operating state when
the temperature decrease below 70◦ C and emit the led.overhead.cleared.
Variables
led.temperature
The led temperature in 0.1 ◦ C unit.
led.temperature.period
The led temperature refresh period in ms. Maximum value is 400ms. This trigger the led.temperature.updated
event.
timers.period
The period of seven general purpose timers. 0 mean disabled. Maximum period is 400ms. The
timerX event expire each time the period timeout.
Events
led.overheat
The LED has overheated, it is temporarly switched off, it will be automatically switched on when
the temperature is sufficiently lower.
led.overheat.cleared
The LED overheat condition is cleared, if the LED was on it is automatically switched on.
led.temperature.updated
The LED temperature is updated. You can read the new temperature in the led.temperature
variable.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives Functions
led.rgb(r,g,b)
Set the LED RGB brightness for each red (r), green (g) and blue (b) componnant. The maximum
value for each componnant (full brighness) is 1000. The minimum value (switched off) is 0.
Settings
There is no settings with this module.
Hidden Variables
There is no specific hidden variables in this module.
6.2.9
gripper-sensor
This aseba node manage the force sensor of the marXbot. The general units are in metric gram.
42
Variables
force.period
The update period of the force sensor, in ms. Max value is 400ms.
force.mag
The magnitude of the force sensor in metric gram.
force.angle.aseba
The angle of the force vector. The unit is aseba angle see section 6.2.1.
force.angle.deg
The angle of the force vector, in degree [0-360].
force.x
The force on the X axis, in metric gram.
force.y
The force on the Y axis, in metric gram.
timers.period
The period of six general purpose timers.
Events
sensor
This even is trigged each time a new force mesurement is available.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives Functions
There is no natives function specific to this module.
Settings
0
X offset
1
Y offset
Hidden Variables
force. raw x
The raw X axis force mesurement without offset cancelation.
force. raw y
The raw Y axis force mesurement without offset cancelation.
potx.tot r
The X axis potentiometer end-to-end total resistance, in 0.1 kOhm.
potx.wpos
The wipers positions of the X axis potentiometers.
poty.tot r
The Y axis potentiometer end-to-end total resistance, in 0.1 kOhm.
poty.wpos
The wipers positions of the Y axis potentiometers.
43
Hidden Natives Functions
potx.set(pot, value)
Set the X axis potentiometer pot wiper position to value.
potx.update()
Read back the X axis potentiometers wiper position.
potx.save(pot)
Save X axis potentiometer pot wipers position in EEPROM.
poty.set(pot, value)
Set the Y axis potentiometer pot wiper position to value.
poty.update()
Read back the Y axis potentiometers wiper position.
poty.save(pot, value)
Save Y axis potentiometer pot wipers position in EEPROM.
6.2.10
griper-led
This node manage the gripper and led ring of the marXbot. Is manage two motors, one for the gripper
fingers, and the other to manage the gripper rotation. The finger motor use a standard motor module
interface named gipper. The speed unit is degree/100ms and the position is in degree. The rotation
motor use a standard motor module interface named rev. The speed unit is degree/10ms and the position
is in degree. Please refer to section 6.2.3 for detailed documentation of both motors.
Variables
timers.period
The period of two general purpose timers.
Events
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives Functions
set led(n,r,g,b)
Set the LED RGB brightness for each red (r), green (g) and blue (b) componnant for the led number
n. The maximum value for each componnant (full brighness) is 63. The minimum value (switched
off) is 0. The led number 0 is the first led. The led number 11 is the last led.
gripper.enc.reset
Reset the gripper motor position. Usefull to set the motor initial (0) position.
rev.enc.reset
Reset the rotation motor position. Usefull to set the motor initial (0) position.
Settings
There is no others settings than the standard motor modules settings.
6.2.11
rab2
This section describe the Range and Bearing module of the marXbot. The Range and Bearing use the
infrared medium to sens the relative position of other surrounding robots and 2.4Ghz radio to synchronize
the swarm and exchange data. The RF and infrared medium access are both TDMA and syncronized
using an unique address on each node and a predetermined network size. This implies the following
restrictions:
44
• Each node of the network must have an unique ID.
• Each node of the network must use the same network size.
• Each node of the network must use the same medium access time.
Variables
rf.id
The ID of the RAB on the network. Must be unique on the network
rf.swarm size
The swarm size. Must be higher than 3 and smaller than 253.
rf.slot time
The medium access time in microsecond. Minimal value: 3000us. Recommanded value: 4000us.
rf.max tx
The number of time the RAB module will send a packet without having any answer. Put 0 to
disable this behavior.
rf.tx data
The data to send. The first 13 bytes will be transmitted (little endian).
rf.ev.node id
The node concerned by the event. See events node.new, node.lost.
tx.power.en
Enable the TX power amplifiers. Only impact infrared medium. If TX is disabled the node will be
visible on the RF network, but not on the infrared.
rx.power.en
Enable the RX power amplifiers. Only impact infrared medium. If RX is disabeld the node will not
be able to sens the others robot position but will be able to communicate over the RF.
rx.source
The received’s ID.
rx.data
The data payload of the received packet. Only the first 13 bytes are valid (little-endian).
rx.range1
The first stage amplifiers value on the IR sensors.
rx.range2
The second stage amplifiers value on the IR sensors.
rx.range3
The third stage amplifiers value on the IR sensors.
rx.dist
The estimated distance of the emiting robot. Units are mm.
rx.angle
The estimated angle of the emiting robot. Units are aseba angle, see section 6.2.1.
timers.period
The period of four general purpose timers.
45
Events
node.new
This event is emitted when a previously never seen node is emiting a packet. See rf.ev.node id.
network.down
This event is emitted when the RAB stop to emit because no answer has been received for a too
long time. See rf.max tx.
network.start
This event is emitted when the RAB is creating a new network.
node.lost
This event is emitted when a node which was previously emiting stopped emiting for more than 255
packets.
packet.rx
This event is emitted when a new packet is received on the RF.
packet.tx
This event is emitted when a new packet is transmitted on the RF.
timerX
Thoses event are trigged each time the corresponding timer expire. See the variables timers.period.
Natives Functions
rf.start()
Start the RF network.
rf.stop()
Stop the RF network.
Settings
0 ... 7
Horizontal IR sensors first stage gain saturation values
8 . . . 15
Horizontal IR sensors second stage gain saturation values
16 . . . 23
Horizontal IR sensors third stage gain saturation values
24 . . . 71
Calibration values
72
The default RAB ID
73
The default network size
74
The default slot time
75
The RF frequency to use
Hidden Variables
rx. dist mm
The estimated source distance on each horizontal IR sensors in millimeters.
46
6.2.12
Magnetic-Gripper-Main
This section describes the Magnetic-Gripper-Main node of the Marxbot. This node lies in the base of
the general module called Magnetic-Gripper. It is bound with the Magnetic-Gripper-Front node which
contains the gripper, some sensors and the RFID tag system.
In the Magnetic-Gripper-Main node, you will find three motors called mot lif t, mot rot and mot tilt
which are responsible respectively of the gripper lifting, the turret rotation and the gripper tilting around
its central axis. The motors are controlled through the standard motor module, see 6.2.3.
Beside the motors, this node has one strain gauge, measuring the lateral bending of the arms.
Variables
Most of this node’s variables concern the motors. For a description of each of these variable, see 6.2.3.
gauge bend.value
This is the value read by the strain gauge measuring the lateral bending of the gripper’s arms. This
value can give a result between 0 and 4095. If correctly calibrated, it should be around 2048 when
the arms are not bent. When the variable displays a value bigger or smaller than 2048 by 1500,
it means that a force of approximately 5000 N is applied laterally to the gripper. The calibration
method used a mass of 500 gramms attached to the gripper to determine the gain factors. This
means that if gauge bend.value displays 3548, an force equivalent to a dead mass of 500 gramms
is pressing laterally on the gripper in one direction and if gauge bend.value displays 548, the same
force is applied in the other direction.
gauge.period
This is the period between two measurement of the strain gauges. It can be set between 1 and
1000. The smaller the period is, the busier the microcontroller will be. A good value is typically 100
(meaning the sensor is updated every 0.1 second).
led[0], led[1]
These leds are not visible unless you demount the robot. They are used in debugging phases.
timers.period[0...3]
These are three timers that you can set to a different period. The maximum period is 400 ms.
Events
Most of the events concern the motors, for more details see 6.2.3.
mot rot index
This event is emitted when the rotation motor reach the index. This happen when the robot is
aligned with its treels, by default it is the position 0 of this motor.
gauge.updated
This event is emitted when the strain gauges are updated.
timer0, timer1, timer2
These events are emitted when the timer 0, 1 or 2 gives a clock count.
Natives Functions
The only three native functions of this node concern the reset of the motors encoder. You can use them
with: call mot lift.enc.reset(), call mot rot.enc.reset() or call mot tilt.enc.reset(). These functions will set
the encoders of the motors to 0.
If you change the reference of the motors, be careful not to go out of the borders the
motors can reach. This could destroy the gripper-arms part.
47
6.2.13
Magnetic-Gripper-Front
This section describes the Magnetic-Gripper-Front node of the Marxbot. This node lies in the gripper of
the module called Magnetic-Gripper. It is bound with the Magnetic-Gripper-Main node.
In the Magnetic-Gripper-Front node, you will find 22 IR sensors, two leds (not visible if the robot is
mounted) and an RFID scanner.
The IR sensors are splitted into four groups. The first ones, IR[0. . . 9], are the frontal sensors. The
second ones, IR[10. . . 19], are the bottom ones. The last two are the left and right sensor, IR[20] and
IR[21]. The values of these sensors vary from 0 (nothing is in front of the sensor) and 4095 (something is
sticked to the sensor).
Variables
ir.period
This is the period between two measurements of the IR sensors. If set to 0, the sensors are deactivated.
If you want to use the microphones, the variable ir.period cannot be set to zero (otherwise, the
microphone are not powered).
ir.ambiant
These 22 values give the ambiant value measured by the IR sensors.
ir.reflected
These 22 values give the reflected value measured by the IR sensors.
ir.delta
These 22 values give the difference between the ambiant and the reflected value of the IR sensors.
These are the value to be used to measure a distance. This will not take account of the ambiant
lightning (as long as the sensor is not saturated).
led[0], led[1]
These leds are not visible unless you unmount the robot. They are used in debugging phases.
rfid.enable
This is a bool value enabling the rfid module (1) or disabling it (0).
timers.period[0...2]
These are two timers that you can set to a different period. The maximum period is 400 ms.
Events
ir
This event is emitted when the IR sensors are updated.
rfid.scan done
This event is trigged when the scanning operation of the rfid chip is over. You can check the result
of scan with the rfid.count() native function. See also the rfid.scan() native function.
rfid.read done
This event is trigged when the read operation on a rfid tag is done. The result of the scan is now
available in the argument data passed to the rfid.read() native function.
rfid.write done
This event is trigged when the write operation on the rfid tag is done. The status (success/failure)
is available in the failed argument passed to the rfid.write() native function.
timer0, timer1
These events are emitted when the timer 0 or 1 gives a clock count.
sound
This event is emitted when the sound buffer of both microphones is filled.
48
Natives Functions
rfid.scan
Start a RFID scan. This will initiate a standard enumeration of all the RFID tags in close contact
with the antenna. A maximum of five tag will be discovered, this should not be a problem as the
antenna can power a maximum of 3 tags. This native function will trigger the rfid.scan done
event when the enumeration process is over.
rfid.count(n)
Get the number of RFID tags found by the last enumeration.
rfid.result(n, id, rssi, block, count)
Get the ID, received signal strength indication, block size and block count of the tag number N.
n Tag number. Must be < to the number returned by rfid.count().
id The 64bits unique identifier of the tag.
rssi Received signal strength indication. Minimum value: 0, maximum value: 31.
block The tag block size in bytes.
count The tag block count.
rfid.read(i, b, d, failed)
Initiate a read operation of the EEPROM block b of the RFID tag with unique identifier i. When
finished the rfid.read done event is generated. If failed is equal to 0, the data read are placed in
the d If failed is nonzero the read has failed, thus d is noninitialised.
Note: The variables failed and d must not be touched between the call to rfid.read() and the
rfid.read done event.
rfid.write(i, b, d, failed)
Initiate a write operation to the EEPROM block b of the RFID tag with unique identifier i. When
finished the rfid.write done event is generated. If failed is equal to 0, the data has been succefully
written. If failed is nonzero the write operation has failed.
Note: The variables failed and d must not be touched between the call to rfid.write() and the
rfid.write done event.
sound.buffer(l[128],r[128])
This function gets the two buffer of the microphones. The left and right buffer is a table of 128
elements. If you take the difference between the maximum and the minimum value of each table,
you will get the volume of the recorded sound.
servo.open()
This function disables the magnetic gripper.
servo.close()
This function enables the magnetic gripper.
6.2.14
Magnetic-Gripper Motors calibration
This section will describe the motors calibration process. Indeed, the tilting and lifting motors must not
go beyond some physical limits illustrated on 6.2. As long as no absolute landmark exists on which these
motors can rely, a calibration process has been established.
The Initial calibration.aesl file will perform this calibration.
1. The robot will turn on itself using the rotation motor. This is the only motor with an absolute index.
As soon as it has reached this index, the robot knows it is in the right position. It will then turn
until it reaches the index and reset the rotation motor encoder to set the zero position in parallel
with the treels.
2. The robot will lower the gripper until it raise itself a bit, then will set its arms a bit higher than the
ground. This manoeuvre is done to detect approximately the ground.
49
Figure 6.2: The boundaries of the lifting and tilting motors.
3. The robot will tilt its gripper around itself (beyond the boundaries) to detect the ground and to be
able to set the tilting motor to its zero. Once done, the gripper returns to its new zero position.
4. The lifting arms will go down again to adjust the ground detection. When it detected the ground,
this time through the IR sensors, the robot is stopped and the three motors are calibrated to their
respective 0.
The robot should lie on a flat, obstacle-less, white surface during the calibration.
If during the phase 2 the robot doesn’t stop lifting itself after a few second, you
should press on the dsPic Reset to stop the calibration process, set the arms to a
position included in the boundaries and restart the calibration process. The lifting
and tilting motors must be in a position included in the boundaries before initiating
the calibration process. If at the end of the phase 3 the gripper is not at its 0 position,
the calibration process must be stopped (dsPic Reset) and restarted.
This calibration process can be ineffective on a few robots (it depends on the IR
values and the motor current feedback which can change from one robot to another). If you prefer, you can place the robot at the desired position yourself and
the launch the three functions call mot lift.enc.reset(), call mot rot.enc.reset() and
call mot tilt.enc.reset() which will set the motors encoders to 0 (the ”zero” position is when the turret is aligned with the treels and when the gripper lies on the
floor with the magnetic gripper at 0◦ as on 6.2). The three motors are reversible,
which means that you can move them by hand. But be careful, sometimes, the
mechanical links between the motor and the part you want to move can be a bit stiff.
The easiest method to place the motor where you want by hand is to set the variables mot lift.pid.enable, mot tilt.pid.enable and mot rot.pid.enable to
1 and then to set mot lift.pid.target current, mot tilt.pid.target current
and mot rot.pid.target current to 0. This means that the motors aimed current
is 0. When you will try to move one of these motors, it will generate a current.
Then, the motor will try to cancel this induced current by powering the motor in the
direction you are pushing. In the end, the motor will help you move itself and the
possible stiffness will disappear.
50
Bibliography
[1] M. Bonani, V. Longchamp, S. Magnenat, P. R´etornaz, D. Burnier, G. Roulet, F. Vaussard, H. Bleuler,
and F. Mondada. The marXbot, a miniature mobile robot opening new perspectives for the collectiverobotic research. In Intelligent Robots and Systems (IROS), 2010 IEEE/RSJ International Conference
on, pages 4187–4193. IEEE.
[2] Michael Bonani. Robotique collective et auto-assemblage. PhD thesis, Lausanne, 2010.
[3] St´ephane Magnenat. Software integration in mobile robotics, a science to scale up machine intelligence.
PhD thesis, Lausanne, 2010.
51
Was this manual useful for you? yes no
Thank you for your participation!

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

Download PDF

advertisement