基于多峰函数优化的无线传感网稀疏事件检测  

MULTIMODAL FUNCTION OPTIMISATION-BASED SPARSE EVENT DETECTION IN WIRELESS SENSOR NETWORKS

作　　者：华涛[1] 李来祥[1] 

机构地区：[1]聊城大学网络信息中心,山东聊城252059

出　　处：《计算机应用与软件》2015年第1期114-117,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(61104179)

摘　　要：为提高无线传感网稀疏事件检测的实时性,提出一种基于多峰函数优化的压缩感知事件检测算法。该算法利用多峰函数最小化方法求解压缩感知检测算法中的欠定线性方程组问题。首先构造连续高斯函数对多峰目标函数进行逼近,而后通过求解连续函数的最优化问题得到多峰函数最小化的解。仿真实验表明新算法与以往的压缩感知检测算法相比,同一长度的事件向量,多峰函数最小化检测可在保证重构准确度的前提下有效缩短检测时间,而重构误差与经典算法基本一致。For improving the real-time performance of sparse event detection in wireless sensor networks （WSN）, a compressed sensing （CS） event detection algorithm based on multimodal optimisation is proposed in this paper. In it the underdetermined linear equations problem in CS detection algorithm is transformed to muhimodal function minimisation algorithm. First the muhimodal object function is approximated by constructing a continuous Gaussian function, and then the solution of multimodal function minimisation is derived by solving the optimisation problem of continuous functions. Simulation experiments shows that in new algorithm, for the event vector with same length, the minimisation detection of muhimodal function can effectively shorten the detection time while ensuring the reconstruction accuracy compared with previous CS detection algorithms, and the reconstruction error are similar to those of the classical algorithms.

关 键 词：无线传感器网络 压缩感知 稀疏事件检测 实时性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]