This thesis presents derivation of a trajectory tracking and stair climbing stabilization controller for a 4x4 skid-steered wheeled mobile robot (SSWR). The robot vehicle is a sturdy platform actuated by DC motors capable of traversing difficult terrain. For trajectory tracking, an essential capability for autonomous operation, a reliable and robust controller is needed. In addition, as the vehicle is unstable with manual control while climbing stairs, the controller is required to stabilise it during stair/ramp climb.
The robot vehicle is modelled with six degrees of freedom (6DOF) rigid body equations and an efficient control algorithm, called Trajectory Linearisation Control (TLC), is used to tackle the challenges posed by nonlinearities of the model. In TLC, state dynamics are linearised along the trajectory being tracked and PI control is used to stabilise tracking error dynamics. Kinematics and dynamics are controlled individually using feedback loops, where the former constitutes the outermost loop.
The main contribution of this work is analysis of 6DOF physical model and a consolidated simple controller for planar tracking and stair climbing stabilization for an SSWR. Simulation results promise that a stable climb on 20° steep staircase is possible with current vehicle configuration. Monte Carlo simulations prove that the controller is robust to realistic dispersions of frictional and physical parameters. Effects of perturbations in these parameters have been studied and improvements in mechanical design are suggested.