基于衰减信道的无线传感器网络次优决策融合算法

Sub-optimal decision fusion under fading channels in wireless sensor networks

无线传感器网络中,节点的本地判决通过衰减信道传输到融合中心,因此融合中心必须结合信道模型做出最优判决。瑞利衰减信道模型下,似然比决策融合规则性能最优,但需对信道参数进行估计,要占用大量资源。本文利用条件概率密度函数推导了新的融合规则,减少了对信道参数的估计,在高信噪比和低信噪比情况下作了进一步近似,提出了新的次优融合规则,并将误差处理相关理论运用到融合算法中。仿真结果表明相比于已有的次优融合规则,本文提出的算法提高了判决性能。

In wireless sensor network, local decisions of nodes are transmitted through fading channels to fusion center where the optimal decision must be made according to the channel models. Under Rayleigh fading channel model, likelihood ratio rule is the optimal method; however channel parameter estimation requires consuming a great deal of system resource. A new fusion rule with less channel parameter estimation is derived using conditional probability density function. Sub-optimal fusion rules without channel parameter estimation are proposed, which are the approximation under low and high signal-to-noise ratios. Error processing theory is applied to the fusion rules. Simulation results show that the proposed fusion rules exhibit better performance compared with the existing sub-optimal rules.