The well-known sensor selection issue is of great significance in wireless sensor network. We aim at selecting a subset of out of sensors to conduct the corresponding hypothesis test based on their observations such that the false alarm rate attains the minimum, which has been proved to be NP hard. In this paper, we relax the original problem to a tractable one and propose a new algorithm, which can achieve the KKT point. Moreover, compared with the existing algorithm, our algorithm is more efficient since it has lower computation complexity. Simulations also illustrate that the global optimum can be reached in most cases.
