Threshold Optimization in Distributed TM-CFAR Adaptive Acquisition Serial Search System Using Particles Swarm Optimization Technique

Abstract

This work investigates the detection threshold optimization of the distributed trimmed-mean constant false alarm rate (TM-CFAR) algorithm. The algorithm TM-CFAR is chosen to solve the serial acquisition of the PN sequences problem. The acquisition system uses several identical sensors; every individual sensor makes a local decision. The overall decision, which is zero or one, is obtained at the data fusion center, which is grounded by “AND” and “OR” fusion rules, in the case where signals are independent from sensor to other. Under Rayleigh fading channel assumption, the analytic expressions of false alarm and detection probabilities are derived. The proposed system generates non-linear multi- variable equations, which are difficult to optimize using conventional optimization methods. To overcome this problem, an efficient methodology for simulation based particles swarm optimization (PSO) is suggested from a variety of meta-heuristic techniques. The obtained results demonstrate that, the proposed optimization method shows a powerful and useful tool to solve such problem in terms of achieving lower false alarm probabilities and higher detection probabilities.