LPI-Based Power and Dwell Time Allocation for Target Tracking via Manifold Optimization

In modern electronic warfare environments, resource allocation can improve the tracking accuracy and anti-interception capability of radar networks. The main challenge is to coordinate the transmit power of each radar to reduce the Posterior Cramér-Rao lower bound (PCRLB) while ensuring low interception. The problem is difficult to solve due to the non-convex nature of the low interception constraint. Existing methods are mainly divided into heuristic methods and convex relaxation methods (CRM), but they are computationally expensive or have large relaxation losses. We note that real sphere manifolds naturally satisfy the transmit power constraint and the dwell time constraint, while the exact penalty term is suitable for handling the low intercept constraint. Using these properties, we propose an exact penalty manifold optimization (EPMO) framework for power and dwell time allocation. First, the exact penalty term is introduced and the real sphere product manifold (RSPM) is constructed to transform the original problem into an unconstrained problem. Then, we derive a parallel limited Broyden-Fletcher-Goldfarb-Shanno (PL-BFGS) algorithm to solve the problem. Compared with existing methods, this method has the following advantages: i) the tracking accuracy is improved by 6% compared to SPG; ii) the computation time is reduced by an order of magnitude compared to PSO and by about 40% compared to SPG.