Sensor Allocation and Quantization Schemes for Multi-Band Cognitive Radio Cooperative Sensing System

We investigate cooperative sensing schemes for a cognitive radio (CR) network operating in multiple primary bands. The CR spectrum sensors use energy detection to identify the presence of the primary signals. We consider a parallel fusion architecture in which all the sensing devices send their binary quantized sensing information to an access point, which then applies a fusion rule to determine the presence of a primary user in each band. In this paper, we first propose schemes to assign sensors to various primary bands. We then study efficient quantization schemes when `OR' fusion rule is used in each band at the access point. We show that the optimal quantization scheme is, in general, non-convex and propose a suboptimal solution based on convex restriction of the original problem. We further study quantization schemes for general k-out-of-N fusion rule. We compare the performance of the proposed schemes using simulations.