无线音频传感器网络中低能耗数据采集方法研究

发布时间：2018-05-04 08:58

本文选题：数据采集 + 随机采样　；参考：《佳木斯大学学报(自然科学版)》2015年05期

【摘要】：计算、存储空间、电能等资源极其有限的无线传感器网络节点,难以实现宽带音频、视频、图像等信号的采集和大流量采样数据的无线传输.针对此问题,结合压缩感知理论,提出了一种基于随机采样的音频信号采集方案.该方案中音频采集节点直接对模拟的音频信号进行低速随机采样,并将少量的低速采样得到的采样值传输至用户端,复杂的信号重建过程和信号分析在用户端完成.在单纯随机采样的基础上,提出了加性的随机采样时刻序列生成方法、推导了随机采样等效观测矩阵,并对随机采样等效观测矩阵的性能进行了仿真分析.仿真实验表明:该方案合理可行,以0.4倍奈奎斯特频率对音频信号采样时,信号重建误差低于0.02;以0.5倍奈奎斯特频率对音频信号采样时,信号重建误差低于0.01.在噪声模型下,该方案有一定抑制噪声功能.该方案在保证信号重建质量的前提下,能够大大降低节点数据采集与传输的能耗,从而有效解决无线音频传感器网络中由于数据量大造成的网络拥塞和大能耗问题.
[Abstract]:Wireless sensor network nodes with limited resources such as computing, storage space and electric energy are difficult to realize wideband audio, video and image acquisition and wireless transmission of large traffic sampled data. In order to solve this problem, an audio signal acquisition scheme based on random sampling is proposed. In this scheme, the audio sampling node directly samples the analog audio signal at low speed random, and transmits the sampling value from a small amount of low speed sampling to the user side. The complex signal reconstruction process and signal analysis are completed at the user side. On the basis of simple random sampling, an additive method of generating random sampling time series is proposed. The equivalent observation matrix of random sampling is derived, and the performance of the equivalent observation matrix of random sampling is simulated and analyzed. The simulation results show that this scheme is reasonable and feasible. When the audio signal is sampled at 0.4 times Nyquist frequency, the signal reconstruction error is less than 0.02, and the signal reconstruction error is less than 0.01 when the audio signal is sampled at 0.5 times Nyquist frequency. Under the noise model, the scheme has a certain noise suppression function. This scheme can greatly reduce the energy consumption of node data acquisition and transmission on the premise of guaranteeing the quality of signal reconstruction, thus effectively solving the problem of network congestion and large energy consumption caused by the large amount of data in wireless audio sensor networks.
【作者单位】：喀什大学物理与电气工程学院;
【基金】：国家自然科学基金青年科学基金项目(11404151)
【分类号】：TP274.2;TP212.9;TN929.5

【相似文献】