Parameter Optimization of Decentralized OS-CFAR System Based Modified PSO Method
For decentralized ordered statistics (OS) constant false alarm ration (CFAR) detection system, the parameter estimation and performance analysis in complicated detection condition is a typical nonlinear optimization problem. Owing to the nonlinear property of distributed OS-CFAR detection system, it is seriously difficult to obtain optimal threshold values using some optimization method at the fusion center. This paper provides a novel solution based on an effective and flexible particle swarm optimization (PSO) algorithm. As a novel evolutionary computation technique, PSO has attracted much attention and wide applications, owing to its simple concept, easy implementation and quick convergence. Using this approach, all system parameters can be optimized simultaneously. The simulation results show that the proposed approach can achieve effective performances with the above method.