Task Modelling in Collective Robotics

  • Authors:

  • C. Ronald Kube

  • Affiliations:

  • Department of Computing Science, University of Alberta, Edmonton, Alberta, Canada T6G 2H1/ E-mail: kube&commat/&commat/cs.ualberta.ca, zhang&commat/cs.ualberta.ca

  • Venue:
  • Autonomous Robots
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

Does coherent collective behaviour require an explicit mechanismof cooperation? In this paper, we demonstrate that a certain class ofcooperative tasks, namely coordinated box manipulation, are possiblewithout explicit communication or cooperation mechanisms. Theapproach relies on subtask decomposition and sensor preprocessing. Aframework is proposed for modelling multi-robot tasks which aredescribed as a series of steps with each step possibly consisting ofsubsteps. Finite state automata theory is used to model steps withstate transitions specified as binary sensing predicates calledperceptual cues. A perceptual cue (Q), whosecomputation is disjoint from the operation of the automata, isprocessed by a 3-level finite state machine called a Q-machine.The model is based on entomological evidence that suggests localstimulus cues are used to regulate a linear series of building actsin nest construction. The approach is designed for a redundant set ofhomogeneous mobile robots, and described is an extension of aprevious system of 5 box-pushing robots to 11 identical transportrobots. Results are presented for a system of physical robots capableof moving a heavy object collectively to an arbitrarily specifiedgoal position. The contribution is a simple task-programming paradigm for mobile multi-robot systems. It is argued that Q-machines andtheir perceptual cues offer a new approach to environment-specifictask modelling in collective robotics.