基于并行离散群居蜘蛛优化算法和压缩感知的WSNs稀疏事件检测

发布时间：2018-06-07 18:09

本文选题：无线传感器网络 + 稀疏事件检测　；参考：《中国电子科学研究院学报》2017年02期

【摘要】：对无线传感器网络(WSNs)弱稀疏性事件检测问题进行研究,提出了一种基于并行离散群居蜘蛛优化算法和压缩感知的WSNs稀疏事件检测方案。该方案采用压缩感知(CS)技术进行稀疏事件分析检测,针对事件向量稀疏度未知的特性,设计基于MPI框架的并行离散群居蜘蛛优化算法(PDSSO),重新定义蜘蛛编码方式和自适应迭代进化机制,给出并行转移策略,并将PDSSO应用于CS重构算法中;针对观测字典难以满足约束等距条件的特点,对稀疏矩阵和测量矩阵进行奇异值预处理操作,在保持稀疏度不变的基础上提高了算法重构性能。仿真结果表明,与GMP等检测方法相比,该方案有效提高了WSNs稀疏事件检测成功率,降低了误检率和漏检率。
[Abstract]:The problem of weak sparsity detection in wireless sensor network (WSNs) is studied, and a WSNs sparse event detection scheme based on parallel discrete group spider optimization algorithm and compressed sensing is proposed. The scheme uses CS (compressed sensing) to detect sparse event analysis, and the design base is designed for the feature of event vector sparsity unknown. In the parallel discrete group spider optimization algorithm (PDSSO) of the MPI framework, the spider coding and adaptive iterative evolution mechanism are redefined, the parallel transfer strategy is given, and the PDSSO is applied to the CS reconstruction algorithm. The singular value preprocessing of the sparse matrix and the measurement matrix is carried out for the characteristics of the observational dictionary which is difficult to meet the constraint isometric conditions. As a result, the performance of the algorithm is improved on the basis of keeping the sparsity unchanged. The simulation results show that, compared with the GMP and other detection methods, the proposed scheme effectively improves the success rate of WSNs sparse event detection, and reduces the false detection rate and the missed detection rate.
【作者单位】：重庆城市职业学院;重庆大学;常州大学;
【基金】：全国教育科学规划重点课题(DBC2012039) 重庆市资助课题(113296)
【分类号】：TN929.5;TP18;TP212.9

【相似文献】