基于小波树模型的CoSaMP压缩感知算法

无线传感网络存在网络带宽限制和传感器节点的能耗问题,实际应用中通常希望可以通过重构算法从采集的少量数据中还原出原始信息,压缩感知理论为上述问题提供了一个解决思路。利用压缩感知理论,对无线传感器网络中温度传感器的监测信号进行了压缩感知的应用研究。针对传统压缩采样匹配追踪(CoSaMP)算法中测量次数多、重构精度低等问题,利用信号的小波系数所形成的连通树的结构特性,提出了基于小波树模型的压缩采样匹配追踪算法。将该算法应用到无线传感器网络监测信号的压缩感知仿真实验中,与传统压缩采样匹配追踪算法的重构性能进行比较,结果表明该算法较传统压缩采样匹配追踪算法在一定范围内对无线传感器网络中的温度信号具有更好的压缩感知性能。

CoSaMP algorithm of compressed sensing based on wavelet tree model

Due to network bandwidth limitation and energy consumption of sensor nodes in wireless sensor networksWSN, people usually want to reconstruct the original signal from little data in practical application. Compressed sensing CS theory has provided a solution for the problem above. Using CS theory, this paper launched the application research on sensed data in WSN. Aiming at the problem of more measurements and lower reconstruction accuracy in traditional CoSaMP algorithm, exploiting the structure characteristic of connected wavelet tree formed by the wavelet coefficients of signal, this paper proposed CoSaMP algorithm based on wavelet tree model. Applied the proposed algorithm to simulation experiments of CS on sensed data in WSN, and compared with the CS performance of traditional CoSaMP algorithm. Comparison results show that the proposed algorithm has a better performance of CS on sensed data in WSN within a certain range.