Feedback-based cooperative primary channel activity estimation for dynamic spectrum access

In dynamic spectrum access networks, spectrum sensing plays a key role in determining the reliability and efficiency of opportunistic spectrum utilization. However, when treating the transmission over a licensed bandwidth as a black box, most energy detection-based sensing methods lack the ability of adapting the sensor to the changes in channel condition. This paper introduces a feedback mechanism into the cooperative sensing framework for estimating the channel SNR and adapting the detector's parameters. The feedback mechanism provides information to local nodes that is used to dynamically control the decision threshold and maintain a stable performance level when channel condition changes. In this framework, the problem of balancing the channel detection opportunity and the interference risk is formulated by an optimization procedure based on a proposed weighted cost function. A queuing model of the primary user activity is employed to predict the channel state and assist grouping the detection samples for estimating the channel characteristics. Such an estimation mechanism requires no complex design and little a priori knowledge of the primary network. Simulation results show that the cooperative scheme improve performance from local decisions and that the introduction of a feedback mechanism results in a notable performance improvement over all schemes.