On Reactive Fuzzy Behavior Systems for Navigation Control of Mobile Robots1

摘要

Many rescue operations in uncertain hazardous environments can be done by using autonomous groundrnvehicles (AGVs). The success of the mission depends on the control system used by the AGV. There are two majorrnmethods of controlling AGVs: the traditional deliberative reasoning systems and the emerging reactive behaviorrnmethods. Due to their responsiveness, there has been an increasing interest in reactive behaviors systems. Fuzzyrnlogic can be used in these reactive behavior systems in order to increase their robustness against sensor andrnenvironmental uncertainties; the resulting system is known as a reactive fuzzy behavior system. Two categories ofrnfuzzy behavior systems have been identified: univalued and multivalued fuzzy behavior systems. This paper discussesrnthe structural differences between these systems and gives a quantitative comparison of their performance by usingrnnumerical simulation results. It is seen that the multivalued fuzzy behavior system performs better than thernunivalued system at the expense of a higher computational load. That is, the multivalued system can drive thernvehicle among obstacles to the intended goal while following a shorter path than the univalued system.