Optimal sensor scheduling for state estimation under limited channel resources

Abstract

In this paper, we investigate the estimation-oriented sensor scheduling problem over industrial cyber-physical systems. Different from the existing works, we consider a more innovative and practical scenario where channel resources are limited. A novel transmission scheduling framework for sensors monitoring multiple sub-processes is proposed, where the most important sensory data will transmitted with higher priority and reliability. For two-system case, we first present an explicit optimal scheduling strategy under the off-line scenario. In order to further improve the performance, we propose an on-line scheduling strategy based on the arrival information feedback. Moreover, we have theoretically proved the feasibility and superiority of our proposed on-line schedule to the optimal off-line one. Besides, in order to reduce the computation complexity, we present sub-optimal algorithms for more complicated multi-system scenario, and derive out the conditions under which the optimal or sub-optimal schedules can be designed explicitly and separately, which conspicuously reduces the computation complexity. Moreover, the theoretical gap between our proposed sub-optimal schedules and optimal schedule can be determined by solving two relaxed problems. Simulations are conducted to demonstrate and verify the correctness and advantages of proposed schedules.