Transcript DesaiMentorshipPrese..

An Implementation of
Artificial Physics Using
AIBO Robots and the Pyro
Programming Environment
Ankur Desai
December 7, 2006
Naval Research Laboratories
Artificial Intelligence Center
4555 Overlook Ave., SW
Washington, DC 20375
Mitchell A. Potter, Ph.D.
Principal Investigator
Evolutionary Robotics
Coevolutionary Models
Representation Issues
Continuous and Embedded Learning
Adaptive Systems Team
Shared lab space
Variety of robotic equipment
No wireless communications
Upcoming anniversary demonstration
Rationale
Divide tasks between multiple robots
Based on natural behaviors
Unified platforms
Purpose
Determine whether AIBO is an
effective platform for artificial
physics
Create Python module to control the
AIBO robots
Artificial Physics
Developed by Spears
and Gordon in 1999
Each robot treated as a
molecule
Gravitational forces
simulated
Artificial Physics
Grid formation
Resource protection
Sony AIBO
Python Robotics
Interpreted language
Platform-blind
High-level control
Testing Design
Straight line accuracy
Turning accuracy
Correct functioning of simulation
No testing necessary
Materials
Software
SWIG
Tekkotsu
Pyro
Seven AIBO robots
Procedures – Python module
Build C library object files
Create SWIG wrapper
Compile wrapper into dynamic library
Procedures – Odometry
Setup
Place AIBO in empty room
Connect to host computer
Send command
Walk 10 meters
Turn 360°
Measure actual motion
Straight Line Results
Accuracy of Straight Line Odometry
11
10
9
8
Distance (m)
Accuracy of Straight Line Odometry Data
Expected Measured Error Measured
Error
(meters) Walking (m) (%) Crawling (m) (%)
10
9.44 -5.6
9.86 -1.4
10
4.7 -53
10.92 9.2
10
6.04 -39.6
5.68 -43.2
10
10.46 4.6
9.14 -8.6
10
5.24 -47.6
7.46 -25.4
10
10.7
7
3.6 -64
10
9.42 -5.8
10.68 6.8
10
4.8 -52
10.42 4.2
10
6.5 -35
7.22 -27.8
10
4.24 -57.6
9.9
-1
7
Walk
6
Crawl
5
4
3
2
1
0
Trial
Turning Results
Accuracy of Turning Odometry
400
350
300
Angle (°)
Accuracy of Turning Odometry Data
Expected Measured
Measured
(degrees) Walking (°) Error (%) Crawling (°) Error (%)
360
220
-38.89
320
-11.11
360
340
-5.56
250
-30.56
360
190
-47.22
230
-36.11
360
390
8.33
170
-52.78
360
360
0
400
11.11
360
330
-8.33
260
-27.78
360
340
-5.56
390
8.33
360
380
5.56
220
-38.89
360
230
-36.11
340
-5.56
360
360
0
190
-47.22
250
Walk
200
Crawl
150
100
50
0
Trial
Conclusion
Python module successful
AIBO is not a suitable platform
Alternate localization techniques
Use of different robotic models
Reflections
Overall positive experience
Delayed security clearance
Limited wireless access
Difficult commute
Acknowledgments
I would like to thank the Adaptive Systems team at Naval
Research Laboratories Artificial Intelligence Center,
especially Mitchell Potter and R. Paul Wiegand, for their
guidance and support throughout this project.
Literature Cited
Blank, D., Meeden, L., & Kumar, D. (2003). Python robotics: An environment for
exploring robotics beyond LEGOs. SIGSCE ’03, 35, 317-3121.
Ikemoto, Y., Hasegawa, Y., Fukuda, T., & Matsuda, K. (2005). Gradual spatial pattern
formation of homogeneous robot group. Information Sciences, 171, 431-445.
Lee, M. (2003). Evolution of behaviors in autonomous robot using artificial neural
network and genetic algorithm. Information Sciences, 155, 43-60.
Oliveira, E., Fischer, K., & Stepankova, O. (1999). Multi-agent systems: Which research
for which applications. Robotics and Autonomous Systems, 27, 91- 106.
Röfer, T., & Jüngel, M. (2003). Fast and robust edge-based localization in the Sony
four-legged robot league. In Polani, D., Browning, B., Bonarini, A., &
Yoshida, K. (Eds.), RoboCup 2003: Robot soccer world cup VII (pp. 262-273).
Berlin: Springer.
Spears, W. M., & Gordon, D. F. (1999). Using artificial physics to control agents. 1999
International Conference on Information Intelligence and Systems, 1999, 281288.
Tira-Thompson, E. J., Halelamien, N. S., Wales, J. J., & Touretzky, D. S. (2004).
Tekkotsu: Cognitive robotics on the Sony AIBO. Proceedings of the Sixth
International Conference on Cognitive Modeling, 6, 390-391.