高速最適センサスケジューリング

抄録

This paper addresses a sensor scheduling problem for large-scale networked sensor systems. Sensor scheduling is required to achieve power saving or to make efficient use of communication channels. The scheduling problem is formulated as a model predictive control problem with single sensor measurement per time. It is assumed that all sensors have the same characteristics and they have state dependent noise. We propose a fast and optimal sensor scheduling algorithm for a class of systems with a sensor characteristics. Computation time of the proposed algorithm is proportional to the number of sensors, and it is independent of the prediction horizon. In addition, a fast sensor scheduling algorithm is presented for a class of systems with a mild condition.