Sensor data scheduling over a packet-dropping network

In this paper, we consider sensor data scheduling over a packet-dropping network. A sensor having sufficient computation capability runs a local Kalman filter and needs to communicate its local state estimate to a remote estimator. However, the sensor has to decide at each time whether to send its local estimate data to the remote estimator or not due to its limited communication energy. We first provide a necessary condition for a sensor data scheduling scheme to be optimal. We then make use of this necessary condition to construct an optimal sensor data scheduling scheme that minimizes the estimation error at the remote estimator side and satisfies the energy constraint at the sensor side.