Experimental Implementation of Tracking Error Elimination for Omnidirectional Wheelchair Using PD-Fuzzy-P Controller

摘要

Wheelchairs have been an active-research problem in the maneuverability of mobile robots over the past decade. This study aims to develop the transportability of disabled people by controlling the motion of their smart wheelchairs. It presents a design of omnidirectional mobile robot where its wheels can be equipped to replace a conventional electric wheelchair in order to achieve flexible and reliable maneuverability. The decentralized algorithm is used for motion control of the omnidirectional robot, which deals with the three independent components of robot’s motion in the body coordinate frame: rotational moving, horizontal moving and vertical moving, which can be controlled separately with separated and different sub-controllers. The simulation analysis of the proposed study has been tested in authors’ previous work and two approaches of artificial intelligent-based controllers (PD-Fuzzy-P and GA-PID controllers) have been built in order to control optimally the maneuverability of such system. The results have showed that the PD-fuzzy-P controller has converged faster than the GA-PID controller and the robot has been able to track successfully the sharp curves’ maneuverability, such as a 90° corner in the squared shape and a U-turn in the rose shape with an error approaching zero. In this paper, the experimental tests of the PD-fuzzy-P controller are carried out on a real robot developed using EV3 Lego kit parts and Vicon motion capturing system is used to capture the robot’s motion in various trajectory scenarios in order to test the effectiveness of the proposed control design.