Robot Behaviour Conflicts: Can Intelligence be Modularised
Mali Amol Dattatraya
April, 1994
Advisor: Amitabha Mukerjee
(amit@iitk.ac.in)
Abstract
In this paper, we examine one of the fundamental assumptions of behaviour-based models: that complex functionalities can be achieved by decomposition into simpler behaviours. In particular we look at the issue of conflicts among robot behaviour modules. The chief contribution of this work is a formal characterisation of temporal cycles in behaviour systems and the development of an algorithm for detecting and avoiding such conflicts. We develop the mechanisms of stimulus specialisation and response generalisation for eliminating conflicts. Thus the probable conflicts can be detected and eliminated before implementation. However, the process of cycle elimination weakens the behaviour structure.
We show how
Reference:
Mali, Amol D.; and Mukerjee, Amitabha; 1994 Robot behaviour conflicts: can intelligence be modularized? Proceedings of the National Conference on Artificial Intelligence, AAAI-94, p.1279-1284 [gzipped PS file, (62149 bytes)] Mali, Amol D. and Amitabha Mukerjee; 1997 Modularity assumptions in situated agency AAAI Workshop on "Deep Blue and Kasparov: The significance for Artificial Intelligence," Rhode Island, AAAI-97, July 1997. [gzipped PS file, (490481 bytes)] Mukerjee, Amitabha; and Mali, Amol D.; 1998 Reactive Robots are like Amnesics : A Comparative Study in Memoryless Behavior IEEE Transactions on System Man and Cybernetics, April 1999[Thesis List] [ME Home Page]