Sensor Scheduling for Multi-Parameter Estimation Under an Energy Constraint
The authors consider a sensor scheduling problem for estimating multiple independent Gaussian random variables under an energy constraint. The sensor measurements are described by a linear observation model; the observation noise is assumed to be Gaussian. They formulate this problem as a stochastic sequential allocation problem. Due to the Gaussian assumption and the linear observation model, this problem is equivalent to a deterministic optimal control problem. They present a greedy algorithm to solve this allocation problem, and derive conditions sufficient to guarantee the optimality of the greedy algorithm. They also present two special cases of this scheduling problem where the greedy algorithm is optimal under weaker conditions.