Real time PSO based adaptive learning type-2 fuzzy logic controller design for the iRobot Create robot

Abstract

Recently, there has been a considerable interest on learning type-2 fuzy logic systems, essentially on how determining the footprint of uncertainties of linguistic variables. In fact, the complexity and difficulty in developing type-2 fuzzy systems can be located at the time of deciding which are the best parameters of membership functions (MFs). In real robot applications, the task of designing a type-2 fuzzy logic controller is complex enough essentially because the presence of many forms of noise and uncertainties, where the robot while navigating has to control many variables. In this paper we present a novel adaptive learning type-2 fuzzy logic controller (FLC) for robot motion planing task. The MFs are tuned instantanously using real time particle swarm optimization technique. The proposed architecture presented good results which were demonstrated on the “iRobot Create” robot.