Real-Time Mamdani-type Fuzzy Maneuvering Control of a Differential-Drive Mobile Robot

Abstract

This paper concerns the real-time application of Mamdani-type Fuzzy Logic Controllers (FLCs) for maneuvering trajectory tracking of a differential-drive mobile robot, the Quanser QBot 2. Based on a recent fuzzy control strategy that allows the robot to move in both forward and backward directions (FB-Controller), a new FLC with sigmoidal and pi-shaped input membership functions is proposed, the FBm-Controller. The aim of this modification is providing smoother movements during the trajectory tracking, improving real-time performance. The proposal is compared with the FB-Controller and F-Controller, where the latter only allows forward movements. Simulations and real-time experiments performed on QBot 2 robot illustrate the advantages of FBm-Controller over the two other approaches.