A comparative study of fuzzy logic controllers for autonomous robots.

Abstract

This paper presents the results from an experimental comparison of a number of fuzzy logic controllers performing mobile robot navigation. An experiment is described which requires the robot to complete a complex, measurable navigational task. The world's first generalised type-2 fuzzy logic controller is compared to a type-1 and a type-2 interval controller. Visual and statistical analyses show that the generalised type-2 fuzzy controller gives a better performance in consistency and smoothness. The impact of this paper comes primarily from the rigorous use of statistical analysis in mobile robot navigation.