Fuzzy-neural based multi-agent strategy for biped motion control

Abstract

In this paper the problem of motion control of biped is considered. We develop a new method based on multi-agent associated Fuzzy-Neural and Adaptive Stochastic Petri Nets strategy. This method deals with organization and coordination aspects in an intelligent modeling of human motion. We propose a cooperative multi-agent model. Based on this model, we develop a control kernel named Ih4COK ( Intelligent Motion Control Kernel ), which consists of a controllor, coordinator and executor of diflerent cycles of motion of biped. When walhng, MCOK receives messages and sends oflers. A decision making of actions (DM) is developed at the supervisor level. The articulator agents partially planifi the motion of nonarticulator agent associated. The system is hybrid and distributed jknctionally. In the conflictual situations of sending or receiving messages we apply a new strategy: The Adaptive Stochastic Petri Nets (ASPN) which deals with all these situations. Cognitive agents communicate with reactive agents (mnarticulator) in order to generate the motion.