基于不平衡扩展模型的火灾信息分布式压缩感知

Distributed compressed fire signal sensing based on unbalance expander

针对无线传感网络数据传输与计算的不均衡而导致部分节点能耗大的问题,首先结合图论中二部图思想,将不平衡扩展模型应用在分布式压缩感知上,并设计出一种与该架构相对应的分布式算法.该算法通过一个列稀疏度确定的稀疏随机二值矩阵决定节点之间是否实现数据传输,从而将传输和计算任务平均分散在各个节点,并利用二阶锥形规划法对融合中心的数据进行重构.最后,在火灾场中利用不平衡扩展模型的分布式压缩感知网络进行仿真实验,并对算法的优越性和网络的节能性作出详细分析.在仿真过程中,通过分析均方误差和信噪比证明所提出的模型不仅在降低节点能耗上有较好的效果,而且在有噪声环境中可以很好地保证信号的重构性能.

In wireless sensor networks, the huge power consumption of part of nodes brings great hardship for various applications, which is caused by the unbalanced data transmission and calculation. To solve this problem, using bipartite graph thought in graph theory, distributed compressive sensing network architecture based on unbalanced expander is proposed. Meanwhile the distributed algorithm corresponding to the architecture is designed. This algorithm decides whether or not to transmit data through a fixed column sparse degree sparse random bipartite matrix, then decentralize the transmission and calculation mission to every node equally and reconstruct the data in fusion center by using second-order cone programming. Finally the distributed compressive sensing network based on unbalanced expander is applied to the fire ground simulation experiment and the superiority of the algorithm and the energy conservation of the network are analyzed in detail. In the process of simulation, through analysis of the mean square error and signal-to-noise ratio, it is proved that the proposed model not only has good effect on reducing nodes＇ energy consumption but also ensures the performance for the signal reconstruction in noisy case.