Optimizing distributed detection thresholds for multistatic radar systems

Abstract

In this paper, we propose an approach for optimizing distributed detection thresholds for multi-static radar systems. Under the Neyman-Pearson criterion, local detection thresholds are optimized using the Particle Swarm Optimization (PSO) algorithm. The local thresholds are optimized to maximize the overall detection probability under the constraint of a given overall false alarm probability. The advantages of PSO include its simplicity, few parameters, and efficient global search. Numerical simulation examples are provided for a radar system comprising two local stations for the distributed detection of targets in Gaussian clutter. The results indicate that the OR rule consistently serves as the optimal fusion rule, and thresholds are optimized flexibly to maintain the overall detection performance despite heterogeneous changes in the signal-to-clutter power ratio (SCR) at local stations. In some cases, local threshold optimization can be omitted without significant reduction in the overall system detection performance.