Decentralized, Massively Parallel Path Planning and its Application to Process Control and Multi Robot Systems

Innovative Applications of Massive Parallelism

Volume

Papers from the 1993 AAAI Spring Symposium

Contents

We have presented a model to reduce process-control problems from the real world to configuration-space [27] to a state space [20]. The process-control problem is then solved as a path planning problem in the state space and the results (the trajectory) translated back as instructions for the controlling mechanisms. The idea is quite similar to the application of fuzzy logic to controllers [40]. Instead of the system being modeled analytically by a set of differential equations, their solutions telling the system what action to take, the fuzzy controller handles these adjustments by a fuzzy rule-based expert system -- a model of human thinking process.