UC San Diego
Optimal Control Of Continuous Time Systems With Quantized Actuators
- Author(s): Prior, Gideon
- et al.
Continuous time plants acted upon by quantized actuators are an important class of systems and arise in numerous control applications in the fields of power electronics and motor drives. This thesis presents a new method for stabilizing these systems by selecting input sequences through the evaluation of an energy related control Lyapunov function. Through this approach multiple performance objectives can be targeted including switching frequency reduction and minimization of torque ripple while maintaining fast transient response and providing stability guarantees. Additionally, by abandoning the heuristics used in popular state of the art methods in favor of the proposed stability based approach, non- optimal inputs can be quickly removed from consideration with little computational investment which provides a significant benefit to real time model predictive control strategies. Due to the relatively simple implementation requirements and the flexibility in setting control objectives, the proposed control theory is potentially well suited for a broad range of applications outside of power electronics and motor drives