(Joint work with Miroslav Pajic)
We consider a wireless networked control system with multiple loops closing over a shared wireless medium. To avoid interferences a centralized scheduler decides which control task accesses the channel at each time step, opportunistically based on the random wireless channel conditions that systems experience. We formulate the problem of designing channel-aware scheduling and transmit power allocation mechanisms that guarantee Lyapunov-like control performance for all tasks in expectation over the channel conditions, while they also minimize the total power expenditures. Exploiting the zero duality gap, optimal variables are obtained by solving at the dual domain either offline, or online based on the observed random channel sequence. Simulations illustrate the power savings of the proposed scheme.