A Distributed Sensor Selection Mechanism for Cooperative Spectrum Sensing

In a cognitive sensor network (CSN), selfish sensors can choose to conduct cooperative spectrum sensing (CSS) or local spectrum sensing (LSS). CSS usually improves sensing performance at the price of delay and energy consumption. A selfish sensor only decides to conduct CSS if it is more profitable than performing LSS. A key issue is how to achieve a desired decision outcome that maximizes spectrum utilization under the requirement of self-interest maximization and primary user (PU) protection. To address this problem, we formulate the interactive deciding of sensors as a noncooperative game, and Nash equilibrium (NE) corresponds to a stable decision outcome that no sensor has an incentive to resist unilaterally. It is shown that the desired decision outcome can be obtained by solving a constrained nonlinear 0-1 programming problem. We derive the desired decision outcome for homogenous sensors, and then develop a sensor selection mechanism for heterogeneous sensors to achieve the desired decision outcome in a distributed fashion. The proposed mechanism is shown to have a relatively low complexity. Computer simulations are carried out to validate the effectiveness of the proposed approaches.