Control strategy selection for autonomous vehicles in a dynamic environment

Autonomous mobile vehicles rely on sensors as a primary perception mechanism. Multiple sensors are used to tolerate certain types of errors and inaccuracy due to system faults and uncertain environment. An autonomous vehicle should be able to detect the dependability of various sensor data and choose the best control strategy based on its internal state and the external environment. This paper proposes a fuzzy rule based control strategy selection approach for mobile navigation in a dynamic real world environment. Fuzzy reasoning Petri nets are used for system modeling and parallel reasoning. The control strategy combining the most dependable sensor data can be chosen quickly based on the proposed approach. Case studies are provided to illustrate and verify the approach.