LPI radar network optimization based on geometrical measurement fusion

The power optimization problem and target assignment are investigated to improve low probability of interception (LPI) in a distributed radar network system. The target geometrical localization error and probability of detection are considered as QoS metrics of the network. We introduce geometrical fusion gain as a metric to fuse information for measuring target localization by multiple radars. The problem is then considered as an optimization challenge based on measurement error covariance ellipses, which satisfy detection and localization accuracy of the network. The LPI problem can be solved with two different objectives for the optimization as it minimizes the maximum power and cuts down on total power. Due to the combinatorial nonconvex and nonlinear nature of the optimization problems, relaxations are considered to make the problems more tractable. They can be efficiently solved by dividing them into two subproblems. Firstly, there is the power allocation in which a framework is proposed to compute minimum required power for a radar assignment scheme by applying a numerical method. Secondly, there is the radar assignment in which a heuristic assignment algorithm is suggested to acquire sufficient radar-to-target assignment based on calculated power. The simulation results show that the proposed algorithms considerably improve LPI performance while detection probability and target localization error constraints can be satisfied.