Abstract: For cooperative spectrum sensing based on data fusion, the detection threshold and the configuration of relay users have a great influence on the total error probability. To avoid the exhaustive search, two fast algorithms for the best threshold and for the best configuration are proposed. According to the monotonousness of the total error probability, the search algorithm based on dichotomy is applied to approach the best threshold step by step. The greedy algorithm selecting one best user step by step until the total error probability doesn’t decrease is also applied. Simulations show the proposed algorithms decrease the search times effectively with a high spectrum sensing performance kept.

Key words: cognitive radio, cooperative sensing, data fusion, total error probability

摘要： 在基于数据融合的协作频谱感知中，检测门限与中继用户是影响总错误率的主要因素。为避免采用复杂度高的全搜索方法，分别提出检测门限快速优化算法以及中继用户快速优化算法。前者依据总错误率的单调性，采用二分法逐步逼近最佳检测门限；后者根据贪婪思想，渐次选择一个最佳中继用户直至总错误率不再降低。仿真结果表明在保证高频谱感知性能的前提下，两者搜索次数均远低于全搜索方法。

关键词: 认知无线电, 协作感知, 数据融合, 总错误率