MODELING AND CONTROL FOR TRAJECTORY TRACKING AND MULTI-ROBOT FORMATION OF NONHOLONOMIC WHEELED MOBILE ROBOTS

Abstract

Wheeled mobile robots (WMRs) are difficult to stabilize and control due to nonholonomic constraints. The complexity of control increases when there is a need to control a group of WMRs in a specific formation. The research described in the thesis concerns the modeling, control and formation of nonholonomic WMRs for trajectory tracking. In particular, this research work proposes novel control approaches for WMR trajectory tracking and formation control of multi-robot system. The kinematic model of WMR is studied, and then various kinematic controllers have been implemented to identify suitable kinematic controller for WMR trajectory tracking using performance based simulation results. The trajectory tracking problem is extended by including the dynamic model of WMR along with the model uncertainties and disturbances. A novel adaptive sliding mode state feedback control law is proposed for trajectory tracking, which includes the integral action and hence it is able to remove the steady state errors and reject the external disturbances. The proposed adaptive dynamic controller uses velocities as input commands, which is more practical and appropriate from the view-point of real-time application. The proposed dynamic state feedback controller requires all the states specially WMR linear and angular velocities. Therefore, the trajectory tracking control problem is addressed again in the context of output feedback control for WMR. The tracking formulation is defined with the high gain observer to estimate the linear and angular velocities. It is shown that using high gain observer and a globally bounded state feedback stabilizing controller, the close-loop system performance can be recovered in the presence of un-modeled dynamics. The formation control problem of multi-robot system is addressed using leader-follower formation approach. The kinematic model of the leader-follower formation is developed in the presence of uncertainties and disturbances. It is followed by an integral terminal sliding mode control for robust formation control and finite-time convergence. The proposed controller eliminates the requirement of leader’s velocity information which increases the reliability of multi-robot system. Obstacle detection and collision avoidance are incorporated to maintain the desired formation in the presence of obstacles. The stability analyses is carried out using Lyapunov stability theory. The performances are verified and validated using time invariant and time varying trajectories.