Adaptive motion control for a differentially driven semi-autonomous wheelchair platform

Abstract

Motorized wheelchairs are systems with changing parameters: rolling friction which depends on the ground surface, center of gravity and even the total mass of the system may vary. This affects the performance of PID controllers normally used for wheelchair velocity control. This paper presents an adaptive controller with online parameter estimation for the motion control of a semi-autonomous robotic wheelchair platform. The proposed controller accepts a second-order trajectory as an input rather then the desired velocity at each moment in time. Mathematical and computer models are developed for the wheelchair and the controller. Four simulation experiments with different initial conditions are demonstrated and analyzed in this work based on a custom developed MATLAB/Simulink® framework for wheelchair dynamic modeling and visualization. Advantages and limitations of the proposed approach are discussed.