Intelligent controller for nonholonomic wheeled mobile robot: A fuzzy path following combination

Abstract

This research article presents a novel solution of controller design for the autonomous path-following maneuver of a nonholonomic wheeled mobile robotic (WMR) system subjected to static and dynamic obstacles. Currently, autonomously operated nonholonomic vehicles, capable of completing tasks avoiding collisions, are deployed for low-profile household applications to sophisticated space missions. The geometrically constrained path-following controller with embedded intelligence for collision avoidance is designed based on a twofold hybrid control law. The hybrid control law combines the feedback linearization controller (FBC) with the fuzzy logic controller (FLC). The feedback linearization controller helps the mobile robot to remain on its desired path, while the fuzzy logic controller helps the mobile robot to avoid critical obstacles around the reference path. It is analytically proven that the proposed hybrid control law converges the robot asymptotically to the desired path when it is safely away from obstacles. Further, when the robot detects an obstacle in its sensing range, obstacle avoidance control law acts locally and it deviates itself from the desired path to avoid collision assuring stability through a bounded tracking control law. Considering bounded sensor measurement uncertainties during obstacle detection, it is observed that the fuzzy logic-based collision-avoidance algorithm behaves in a robust manner. The hybrid controller is designed and tested in a simulated environment with the real-life parameters of a nonholonomic wheeled mobile robot. The robot is capable of traversing through the geometrically constrained desired paths avoiding obstacles in an improved manner compared to the existing method. Moreover, the robot is also capable of returning back to its original path when obstacles are safely away. The performance of the novel hybrid control strategy is illustrated for several geometrically constrained paths via simulation to show its effectiveness in terms of fulfilling dual objectives of path following and simultaneous collision avoidance in an improved manner.