Behavior Description and Environment Modelling for Wheeled Mobile Robots of Home Service

Wheeled mobile robots have promising applications in home service. However, the complexity and variety of home environments limited the applications. A novel approach, which is based on automaton, fuzzy logic and neural networks, for behavior description of home robot is proposed. It adopted formal language and automata to describe service behavior and motion behavior simultaneously, and adopted recurrent neural networks to model the environment´s knowledge and to implement a navigator. The network can construct itself on-line automatically.