In spectrum sensing for cognitive radio networks, the secondary users should detect whether the primary user is present as soon as possible to improve the throughput of the secondary system. This paper investigates the problem of minimizing the detection time of log-likelihood ratio test based cooperative spectrum sensing under the constraints on accuracy metrics. We show that the minimum detection time can be achieved by choosing an optimum sample number. Closed-form expressions for the minimum detection time as well as the optimum sample number are deduced. Simulation results are provided to verify the validity of these expressions.
展开▼
