A HIERARCHICAL DESIGN APPROACH FOR INTERVAL TYPE-2 FUZZY CONTROLLERS APPLIED TO MOBILE ROBOTS

摘要

This paper presents a novel hierarchical design approach for interval type-2 fuzzy logic controllers applied to mobile robots. The approach achieves an optimum interval type-2 fuzzy controller when uncertainty is present in the control loop, which is associated with the use of non-deterministic models in the design process. The method consists of three steps: The first one is intended to obtain a stable interval type-2 fuzzy controller as an initial design. The second one optimizes the control system response without uncertainty by using a type-1 fuzzy controller derived from type-2 design. The third one optimizes the control system behaviour through an interval type-2 version of the type-1 fuzzy controller regarding the uncertainty sources affecting the control loop. The method is successfully validated in two different mobile robots. Simulation results for both problems show that the interval type-2 fuzzy controllers designed using this approach always exhibit better performance than optimal type-1 fuzzy controllers when different kinds of uncertainties affect the robot platform.