Sliding Mode Control Design for Two-wheeled Mobile Robots

This thesis not only concerns the development of sliding mode control (SMC) design for two-wheeled mobile robot (TWMR) systems in a rigorous mathematical manner but also focuses on the application of the developed theoretical SMC algorithms in the practical TWMR systems. The significant contents involve trajectory tracking control on a TWMR with caster wheels and setpoint regulation controls for a two-wheeled inverted pendulum (TWIP). For trajectory tracking SMC of the TWMR system, it is assumed that all the system states are accessible for design. In contrast, both full states and partial states are allowed to be accessible for the controls of a TWIP system. The main achievements in this thesis are summarised as follows. • The kinematic system is considered with matched and unmatched uncertainty in the trajectory tracking control of a TWMR system. A new structure of the sliding functions is proposed to help derive the reduced-order sliding mode dynamic, which reduces conservatism in the stability analysis. In the presence of both matched and unmatched uncertainty, the proposed SMC can track the predefined trajectories effectively and robustly. • A conventional SMC, based on a regular-form approach, is developed for the TWIP system under assumption that all system state variables are accessible. The bounds on both matched and unmatched uncertainties are assumed as known functions used in the SMC design to reject uncertainties and improve robustness. Compared with previous work that used constant or linear bounds on the uncertainties, the developed results allow more general nonlinear forms for the bounds on the uncertainties. As a result, the obtained results can tolerate a broader range of uncertainties. • A static output feedback SMC scheme is proposed to regulate the TWIP system when only partial state information is available. Both the stabilisation and setpoint regulation control problems of the TWIP system are addressed. A novel method is introduced to select the feedback gains for regulating the TWIP system intuitively. • A self-developed real-time operating system (RTOS) based software architecture is implemented for the practical TWIP platform to improve the system performance. Moreover, the proposed SMC laws are demonstrated in simulation and on a practical TWMR with passive wheels for trajectory tracking control and a TWIP platform for setpoint regulation control subject to the matched and unmatched uncertainties. The results show the effectiveness and robustness of the designed control schemes when implemented in practical TWIP systems. The simulations of the SMC approaches mentioned above are conducted using Matlab and Simulink tools. Moreover, the associated experimental verifications of the trajectory tracking and setpoint regulation controls are demonstrated on two different TWMR platforms assembled based on the ARM Cortex-M series microcontroller boards as the primary central processing unit.