摘要
利用压缩感知理论重构网络数据可有效减少无线传感器网络数据传输能耗。已有重构算法复杂度高,难以满足无线传感器网络的高实时性要求。为提高基于压缩感知理论的网络数据重构的实时性,提出一种零范数最小化重构方法。首先构造连续函数对离散的零范数函数进行逼近,然后通过求解连续函数的最优化问题得到零范数最小化的近似解。与以往的压缩感知重构方法相比,零范数最小化重构在保证重构准确度的前提下有效减小了算法复杂度。仿真实验验证了所提算法的正确性和有效性。
Utilising the compressed sensing （CS） theory in network data reconstruction can effectively reduce the energy consumption during data transmission in wireless sensor networks （WSN）. However, because of the high complexity, current reconstruction algorithms can not satisfy real-time requirement of WSN. For improving the real-time performance of CS-based network data reconstruction, we propose in this paper a network data reconstruction algorithm which is based on zero-norm minimisation. First, a suitable continuous function is constructed to approach the discontinuous zero-norm function, and then by solving the continuous function optimisation, the approximate solution of zero-norm minimisation is derived. Compared with previous algorithm of compressed sensing reconstruction, zero-norm minimisation reconstruction effectively reduces the complexity of the algorithm while ensuring the accuracY of reconstruction. Simulation experiment verifies the correctness and validity of the algorithm.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第8期253-257,共5页
Computer Applications and Software
关键词
无线传感器网络
压缩感知
零范数
最小化
实时性
Wireless sensor networks Compressed sensing Zero-norm Minimisation Real-time performance
