Optimization of Double Threshold Based Energy Detection in Cooperative Cognitive Radio Network

In this work, we investigate a double threshold based cooperative spectrum sensing scenario. Our objective is to determine the optimal threshold for majority rule which must be selected for minimum error in final decision. The CR sensors are assumed to make local hard decisions based on conventional energy detection technique and communicate one bit decision information to the fusion center. Here we assume that sensors whose test statistics fall in ambiguity region do not report to the fusion center. A majority rule is applied at the fusion center in which at least a threshold number of local sensor decision must favor for presence of primary user (PU) to make the final decision on presence of PU. Since choice of threshold decides error in final decision, we formulate an expression to compute optimal value of threshold of majority rule , which minimizes error in final decision. Further, due to uncertainty in number of sensors with test statistics in ambiguity region, the threshold also becomes a random variable. Hence we derive a statistical model to characterize the density function of number of sensors with test statistics in ambiguity region, and later exploit it to derive an expression for expected value of optimum value of integer threshold. Our simulation results validate our approach in which we show by selecting optimum threshold value for majority rule, the error in final decision is at its minimum value. Moreover, we have also found out optimal values of ambiguity region width for minimum error in final decision using genetic algorithm