多跳筛选无线传感器网络决策融合

Fusion of decisions in multi-hop censoring wireless sensor networks

为了在保证传感器工作效果的同时控制系统能耗,在资源受限的无线传感器网络决策融合系统中同时引入传感器信息的多跳传输方式和筛选传输策略。在瑞利分布信道模型下,推导出该系统的最优融合规则。但该规则需要信道的实时状态信息,实用性差。为此提出基于信道统计分布信息的次优似然比融合规则,并给出了小信道信噪比情况下的简化形式。理论分析和仿真表明,次优规则性能损失较小,运算量更低,信息需要更少,符合资源受限系统的要求;同时使用筛选策略,降低了系统的传输能耗。

In order to guarantee a better operation performance for sensors while controlling the system energy consumption,both multi-hop transmission mode and censoring transmission strategy are introduced into the decision fusion of the resource-constrained wireless sensor networks.The optimal fusion rule is derived under the assumption of Rayleigh channel model.However,the optimal fusion rule requires the real-time channel information which is too costly for resource-constrained sensor networks,so a suboptimal likelihood ratio test fusion rule based on the knowledge of channel statistics is proposed.In addition,the simplified form of the suboptimal fusion rule is also given under low channel signal-to-noise ratio（SNR）.Theoretical analysis and simulations show that the suboptimal fusion rules need much less computation and information,yet exhibits only slight performance degradation.The introduction of censoring strategy also reduces the energy consumption of system transmission.