On the BEP walls for soft decision based cooperative sensing in cognitive radios

Abstract

Cooperative sensing (CS) is important for efficiently acquiring spectrum awareness in cognitive radio systems. The main focus of this paper is on the BEP (bit error probability) walls for soft decision (SD) based CS in cognitive radios. The BEP wall is a performance limitation for CS caused by imperfect reporting channels. In this paper, a distributed detection system using parallel sensing topology and a fusion center (FC) is considered. Each secondary user (SU) sends energy-based multibit SD to the FC over a reporting channel with a certain BEP. The maximum output entropy (MOE) quantization is considered for quantizing the estimated energy. For the considered case, it is established that the BEP wall values for SD based CS with MOE quantization are identical for arbitrary number of bits for quantization. Moreover, the BEP wall values for SD based CS with MOE quantization are proven to be the same as the BEP wall values for the Chair-Varshney fusion rule, which is an optimal fusion rule for one-bit hard decision based CS. Later relation between the BEP wall values of the SD based CS and the K-out-of-N fusion rules is also established.