Opportunistic sensor scheduling in wireless control systems

Konstantinos Gatsis, Miroslav Pajic, Alejandro Ribeiro, George J. Pappas

Abstract: We consider a wireless control system where multiple power-constrained sensors transmit plant output measurements to a controller over a shared wireless medium. A centralized scheduler, situated at the controller, grants channel access to a single sensor on each time step. Given plant and controller dynamics, we design scheduling and transmit power policies that adapt opportunistically to the random wireless channel conditions experienced by the sensors. The objective is to obtain a stable system, by minimizing the expected decrease rate of a given Lyapunov function, while respecting the sensors' power constraints. We develop an online optimization algorithm based on the random channel sequence observed during execution which converges almost surely to the optimal protocol design.