Autonomous Mobile Robotics
During 1990-92, John was with the
Intelligent Robotics Research Centre
at Monash University. With GIRD funding and with BHP as the industrial partner, a small project team group developed a demonstration robot which could navigate through a room containing unknown obstacles. The robot relied on its own vision system to sense its immediate environment. It was able to re-plan its path when an obstacle was encountered.
Technical Summary
|
Photograph courtesy of Monash University
|
The mobile robot platform was a Robuter, which included its own processor for low-level motion-control. Additional on-board processing capacity was provided by a PC and plug-in card with 4 transputers and interfaces to various periheral devices: a radio modem, an ultrasonic localiser, a rangefinder providing 2 1/2 D vision, and a pair of floor-line rangefinders for wall-detection. The radio modem linked the robot to a Silicon Graphics workstation, used for high-level path-planning and monitoring of the executed path.
|
The robot was demonstrated at the 12th International Joint Conference on Artificial Intelligence in 1991.
John's Contributions
- Design of a 4-port interface between transputer ports and RS232C serial lines.
- Design and coding of the motion-control algorithm, using Cornu spirals to account for the differential acceleration of innner and outer wheels during turns. In order to allow the planned motion to be interrupted gracefully when an obstacle was encountered, pre-planned moves were queued by the motion-control transputer and sent to the Robuter in 180 ms increments.
- Design and coding of a polygon-merging algorithm, used to map obstacles encountered during the robot's travels, and to calculate the region occupied by the robot while turning.
- Co-ordination of the software development, particularly the integration of 30 concurrent processes running on 4 transputers.
- Design of the mechanical superstructure of the robot, including a track and electrically-driven carriage, allowing the main rangefinder to observe the environment in any direction.
Skills
- Microcontroller-based design and assembly-language programming
- C programming, including concurrent processes
- Co-ordinate geometry and machine dynamics
- Mechanical design
- Software co-ordination
Reference
J.M. Badcock, J.A. Dun, K. Ajay, L. Kleeman and R.A. Jarvis, "An autonomous robot navigation system - integrating environmental mapping, path planning, localisation and motion control",
Robotica
Vol 11, pp 97-103, 1993.
Back to experience summary page
Back to home page