Sensing Weight Selection using the Genetic Algorithm to Enhance Cooperative Decision

A cognitive radio network (CRN) is an advanced wireless network setup that enhances network efficiency by allowing opportunistic secondary users (SUs) to use free spectrum holes of the primary user (PUs). In CRN, spectrum sensing is of high interest that cannot be achieved by a single user due to the multi-path and shadowing effects. Therefore, more than one users take the sensing responsibility cooperatively to better estimate the PU channel. The fusion center (FC) receives local sensing statistics reported by all cooperative users and combines them to reach a common global decision. Since the SUs located distance apart experiences different channel conditions, therefore, it is necessary to deal with the incoming data received from these SUs differently. The proposed scheme intelligently tries to find optimum weighting coefficients against cooperative user's information and utilize these weights in the global decision of the soft decision fusion (SDF). The genetic algorithm (GA) scheme proposed in the paper can find reliable sensing weights against the users sensing data that produces minimum false alarm with high detection probability and reduce the error probability. The simulations performed at different levels of the signal-to-noise-ratios (SNRs) and a total number of cooperative users result in better sensing response for the proposed GA scheme against maximum gain combination (MGC), Kullback Leibler (KL), and particle swarm optimization (PSO) schemes.