Switching linear parameter varying control for a robotic manipulator

This paper introduces a synthesis study of a switching Linear Parameter Varying (LPV) controller with multiple polytopic subregions, considering separate parameter-dependent Lyapunov functions. The parameter space setup is designed with overlapping subregions in which safe controller switching is ensured by hysteresis switching logic. Robotic manipulators, like any mechatronic systems, are burdened with unmodelled characteristics such as friction phenomena and actuator dynamics, as well as nonlinear dynamics. These effects might be countered by LPV controllers, although they may be too conservative for production. The switching controller can decrease conservatism utilising separate Lyapunov functions while maintaining essential features of a general H∞ LPV controller, such as global asymptotic stability over the entire parameter space. The suggested method is applied on a real-life 2-DOF SCARA-type robotic manipulator, and extensive testing results are presented along with some of the key challenges that arise when the application is moved from simulation to real-life. Primary considerations with respect to the application of switching control are studied in detail, including the number and shape of subregions, overlapping percentage, switching surface orientation, and scheduling parameter selection.