Abstract:Resource allocation for enhancing the target capacity of multiple target tracking (MTT) with desired accuracies for given transmit power is the key issue in radar networks. Most existing methods solve this problem with heuristic evolutionary methods or convex relaxation methods with high computational cost, which lack the real-time adaptability for dynamic threat scenarios. To overcome this issue, we propose a real-time Adaptive Manifold Optimization (AMO) framework. This is achieved by utilizing the inherent real oblique characteristic of the power matrix constraints. Specifically, we construct a real oblique manifold that satisfies the constraints, enabling the problem to be rephrased as an unconstrained problem over the manifold space. Then, we derive a conjugate gradient algorithm for direct optimization of the problem. Simulation results demonstrate that the proposed method outperforms existing approaches in terms of target capacity, MTT accuracy and computational cost.

Radar Resource Allocation for Tracking Target Capacity Maximization Via Manifold Optimization