An antenna subset selection algorithm in distributed MIMO radar for target localization via convolutional neural network

Since the distributed MIMO radar (DMR) has widely spread transmitters and receivers, it can provide higher target detection probability, as well as superior target tracking and localization performance than the monostatic/bistatic radar systems. An effective radar resource allocation scheme can optimise the DMR system parameter and obtain better system performance. In this paper, a critical but limited system resource is optimised, that is, select a subset of the active radar antennas. In this scenario, the convolutional neural network of the antenna subset selection for target localization (LCNN-ASS) algorithm in the DMR is proposed based on two free switching policies. The proposed algorithm is immune to the failure of a single policy and selects antenna subsets from the entire sets with a remarkable computational speed. Therefore, the proposed algorithm increases the flexibility of resource scheduling over traditional algorithms. Simulation experiments and performance analysis demonstrate the localization performance and flexibility of the LCNN-ASS algorithm.