Joint node selection and power allocation for multitarget tracking in decentralized radar networks

Networked radar systems have shown significant advances in target tracking. Reasonable power allocation strategy can sufficiently utilize the limited power resources, hence leading to the improvement of tracking performance. However, towards the existing power allocation strategies, the system configuration is only restricted to centralized architectures. Besides, practical communication requirements and system robustness have not been considered. To tackle these problems, we propose a joint selection and power allocation (JSPA) strategy for target tracking in decentralized radar networks. The optimal fusion is presented to obtain the global posterior estimates in terms of the local filtering densities. Then the corresponding weights for fusion estimation can be updated according to distributed particle filter (DPF). Finally, the decentralized posterior Cramér-Rao lower bound (PCRLB) is derived, and consequently, employed as an optimization metric for JSPA strategy. We also present an effective method to solve it. Numerical results demonstrate the superior performance of the proposed strategy and method.