Antenna subset selection in distributed multiple-radar architectures: A knapsack problem formulation

In this paper, a performance driven resource allocation scheme for target localization in multiple radar systems is proposed and evaluated. An optimal subset of active antennas of predetermined size, K, is selected such that the localization mean-square error (MSE) is minimized. The problem is formulated in a combinatorial optimization framework as a knapsack problem (KP). The Cramer-Rao bound (CRB) is used as a performance metric. Cost parameters, representing operational cost or any other utilization constraints, are associated with each of the antennas. These are incorporated into the KP formulation, integrating decision making factors in the selection process. Antenna subset selection is implemented through an approximation algorithm, by successively selecting antennas so as to maximize the temporal Fisher information matrix (FIM) for a given subset size. The proposed approximation algorithm offers considerable reduction in computational complexity when compared with exhaustive search, supporting distributive processing and low performance loss.