Multichannel cooperative spectrum sensing that integrates channel decoding with fusion-based decision

Abstract

This paper considers coded multichannel cooperative spectrum sensing (MC-CSS) employing local Neyman-Pearson testing at each sensor and channel decoding integrated with fusion-based decision. A joint multichannel decoding and decision fusion (JMCDDF) algorithm for performing a likelihood ratio test at the fusion center is derived based on the belief propagation technique. For benchmark comparison, we also derive the analytical performance of MC-CSS with error-free reporting channels that do not require channel coding, employing equal decision thresholds at each sensor. Using the JMCDDF algorithm, we compare the performance between uncoded and coded MC-CSS schemes when applying low-density parity-check (LDPC) codes in the presence of reporting channel errors. Monte- Carlo simulation results show considerable performance gains when using the proposed coded MC-CSS schemes. A 3-dB saving in link budget can be achieved by such coded MC-CSS schemes with a short (3, 6) regular LDPC code of codeword length n c = 200, over uncoded MC-CSS. Finally, it is shown that in some cases, protecting the primary user (PU) channels by using variable nodes of higher degrees improves the performance. To equalize the performance for all sensed PU channels, we introduce a simple permutation technique, where, in each transmission, different variable nodes are used to protect local decisions.