Dynamic programming formulation of periodic event-triggered control: Performance Guarantees and co-design

While potential benefits of choosing the transmissions times in a networked control system based on state or event information have been advocated in the literature, few general methods are available that guarantee closed-loop improvements over traditional periodic transmission strategies. In this paper, we propose event-triggered controllers that guarantee better quadratic discounted cost performance than periodic control strategies using the same average transmission rate. Moreover, we show that the performance of a method in the line of previous Lyapunov based approaches is within a multiplicative factor of periodic control performance, while using less transmissions. Our approach is based on a dynamic programming formulation for the co-design problem of choosing both transmission decisions and control inputs in the context of periodic event-triggered control for linear systems. A numerical example illustrates the advantages of the proposed method over traditional periodic control.