TUTORIAL T6



Intelligent Multi-Agent Collaborative Systems (IMACS)
Lecturers: Gourab Sen Gupta and Chris Messom, Massey University, New Zealand

Monday, 17 May 2004, 2:00PM-6:00PM

Gourab Sen Gupta Chris Messom


Intelligent Multi-Agent Collaborative Systems (IMACS) is an emerging and rapidly evolving high-technology area related to such fields as Information Engineering, Computer Science, Mechatronics, Robotics and Intelligent Control, Computer Systems Engineering, etc., and enabled by Instrumentation and Measurement techniques and tools. Intertwined with these areas are such disciplines as image processing, wireless data communication, embedded micro-controllers, path planning, obstacle avoidance, fault tolerance, artificial intelligence and "behaviour programming".

In this tutorial, the hardware and software design solutions, theory and engineering techniques related to the intelligent multi-agent collaborative systems will be introduced and discussed. The presentation will employ Robot Soccer System as an integrated platform for multi-disciplinary research on IMACS. Particular attention will be given to the effective data processing, control techniques and strategies. State Transition Based Control (STBC) is a proven technique to implement "behaviour programming", elements of which will be detailed using several case studies. Vision Sensor Noise corrupts planning, prediction and agent behaviour. Kalman Filtering has been effectively used to overcome this problem and the research findings in this area will be summarised. Lastly, the use of the Robot Soccer System as a teaching platform and for robotic competitions will be discussed.

Outline:
  • Cooperative Mobile Robotics - Antecedents and future directions
  • Design of a micro-robot - fast, furious and accurate
  • Robust RF wireless communication
  • Vision Processing for Real-Time Object Identification
  • Managing Complexity in Behaviour Programming using State Transition Based Controllers
  • Robot Soccer System
  • Biped Robot Control
  • Using filtering to improve predictive control and system performance
  • Robot Collaboration and Cooperation
  • Evolving Intelligent Controllers using Genetic Programming.
  • Robot Soccer in teaching