Networked radar systems have been demonstrated to offer enhanced target tracking capabilities. An effective radar resource allocation strategy can efficiently optimize system parameters, leading to performance enhancements. In this paper, two critical but limited system resources are considered for optimization: the number of radar nodes and the transmitted power. In this scenario, a joint node selection and power allocation (JSPA) strategy is developed with the objective of tracking multiple targets. The proposed mechanism implements the optimal resource allocation based on the feedback information in the tracking recursion cycle in order to improve the worst-case tracking accuracy with multiple targets. The network architecture considered in this paper is decentralized so that communication requirements may be reduced while maintaining system robustness. Since the predicted conditional Cram\'{e}r-Rao lower bound (PC-CRLB) provides a lower bound on the accuracy of the target state estimates conditional on the actual measurement realizations, it is more accurate than the standard posterior CRLB (PCRLB) and is thus derived and used as an optimization criterion for the JSPA strategy. It is shown that the optimal JSPA is a two-variable nonconvex optimization problem. We propose an efficient two-step semidefinite programming (SDP)-based solution to solve this problem. Numerical results demonstrate the superior performance of the proposed strategy and the effectiveness of the proposed solution.

distributed multiple-input multiple-output (MIMO) radar system is capable of tracking multiple targets under the "defocused transmit-focused receive" (DTFR) operating mode, in which each transmitter forms a completely defocused beam to illuminate the whole surveillance region and each receiver adopts a focused beam to attain a higher resolution. However, when operating in this mode, there exists a resource optimization problem on the allocation of receive-beams. To address this problem, a receive-beam resource allocation (RBRA) strategy is proposed in this paper. The key mechanism is to implement the optimal allocation between receive-beams and targets based onthe feedback information in the tracking recursion cycle, with the objective of improving the worst tracking accuracy with multiple targets. Since the posterior Cramer-Rao lower bound (PCRLB) provides a lower bound on the accuracy of target state estimates, it is derived and adopted as an optimization criterion. It is shown that the optimal RBRA is a multidimensional nonconvex assignment problem that is NP-hard. We propose an efficient convex relaxation optimization to solve it. Numerical results demonstrate the superior performance of the proposed strategy in terms of the worst-case tracking root mean-square errors