URC 97125 Fuzzy Behavior - based Navigation for Planetary

A b s t r a c t Adaptive behavioral capabilities are necessary for robust rover navigation in unstructured and partiallymapped environments. A control approach is described which exploits the approximate reasoning capability of fuzzy logic to produce adaptive motion behavior. In particular, a behavior-baaed architecture for hierarchical fuzzy control of microrovers is presented. Its structure is described, as well w mechanisms of control decision-making which give rise to adaptive behavior. Control decisions for local navigation result from a consensus of recommendations offered only by behaviors that are applicable to current situations. Simulation predicts the navigation performance on a microrover in simplified Mars-analog terrain.