Adaptive sensing scheduling for cognitive radio systems

Abstract

In cognitive radio (CR) systems, the efficiency of primary user (PU) detection scheme directly affects the system performance. In this paper, we propose an adaptive sensing scheduling scheme for CR networks with a central controller, which is based on the partially observable Markov decision process (POMDP) framework. With the proposed scheme, the CR system detects the PU activation by using one or more energy detections followed by zero or one feature detection, and protects the PU by using the channel switching. In the proposed scheme, the energy/feature detections and the channel switching are adaptively scheduled according to the number of active CR nodes, the received signal power from a PU, and the degree of noise uncertainty. Simulation results show that the proposed scheme provides very high maximum channel utilization to CR nodes while protecting the PU effectively, under time-varying network environment.