多跳无线传感器网络决策融合  

Decision fusion in multihop wireless sensor networks

作　　者：袁晓光[1] 杨万海[1] 史林[1] 

机构地区：[1]西安电子科技大学电子工程学院,陕西西安710071

出　　处：《系统工程与电子技术》2009年第4期976-981,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金资助课题(60472085)

摘　　要：研究了衰减信道下多跳无线传感器网络的决策融合规则,推导出了瑞利分布的信道模型下性能最优的似然比融合规则。该规则需要预知信道的实时状态信息,计算复杂,不适合资源受限的无线传感器网络。为此,提出了仅需要预知信道的统计分布信息的次优似然比融合规则,并分别在大信噪比和小信噪比情况下推导出了该规则的简化形式。理论分析和实验表明,次优规则性能损失轻微,显著降低运算量和信息消耗,适合在资源受限的多跳无线传感器网络决策融合系统中使用。Decision fusion rules for multihop wireless sensor networks in fading channel are investigated. Optimal likelihood ratio rule has been presented under the assumption of Rayleigh channel model. However, acquiring real-time channel information may be too costly for resource constrained sensor networks. Suboptimal likelihood ratio fusion rule is proposed which requires only the statistical knowledge of channel. In addition, when the channel signal-to-noise ratio （SNR） is high and low,the reduced forms are also derived respectively. Theoretical analysis and simulations show that suboptimal fusion rule needs much less computation and information, yet exhibits only slight performance degradation. Suboptimal fusion rule is practicable for decision fusion system in resource constrained multihop wireless sensor networks.

关 键 词：多跳 无线传感器网络 融合规则 资源受限 信噪比 

分 类 号：TN929.5[电子电信—通信与信息系统] TP212[电子电信—信息与通信工程]