压缩感知在通信中的应用研究

发布时间：2018-01-08 03:29

本文关键词：压缩感知在通信中的应用研究　出处：《杭州电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：压缩感知（CS）理论指出，信号可以低于奈奎斯特速率进行采样。采样频率可以不受信号最高频率的限制，而只与信号自身的结构相关。信号在满足变换域是稀疏的或信号是可压缩的条件下，，就不需要遵循传统的先采样后压缩的模式，而是在采样的同时实现数据的压缩，这样就只需采样少量数据，极大地解放了采样器件和存储设备。随着通信技术的发展，宽带通信以及高频应用给采样设备带来了巨大的挑战，因此本文主要研究压缩感知在通信中的应用。本文首先研究基于信号稀疏表示的调制样式识别。由于传统基于小波变换的调制识别算法需要码元同步和较长数据，其计算量较大、耗时严重，难以满足战场通信条件下实时性的要求，针对这些问题本文提出利用小波变换提取信号突变点信息，根据归一化前后信号小波系数稀疏性的不同特点，采用信号小波变换后的稀疏性作为分类识别特征参数，优化了分类判决过程。该算法无需码元同步，算法复杂度低，且在低信噪比条件下，本算法比传统基于小波变换的调制识别算法具有更高的正确识别率。然后研究基于压缩感知的宽带频谱感知方法。当前基于压缩感知的频谱感知研究中通常以频谱稀疏度已知作为先验信息，而实际频谱感知中信道稀疏度是未知且时变的。针对以上，问题本文提出一种稀疏度自适应的宽带频谱感知算法。利用分布式压缩感知和RIP性质对稀疏度进行预估计，通过置信系数更新估计得到频谱支撑集，仿真结果表明本算法在低信噪比条件下性能优于稀疏度已知的频谱感知算法。最后研究基于1-bit的宽带压缩频谱感知。分布式频谱感知网络中节点数据传输占用一定的网络通信带宽，本文提出基于分布式的1-bit压缩频谱感知算法，对节点感知数据进行1-bit量化，仅保留符号信息，简化了数据传输和存储，减轻了网络带宽压力。仿真结果表明该算法是有效的。
[Abstract]:Compression sensing theory indicates that the signal can be sampled at a rate lower than Nyquist rate, and the sampling frequency can not be limited by the maximum frequency of the signal. The signal is only related to the structure of the signal itself. When the signal is sparse in the transform domain or compressible, it is not necessary to follow the traditional mode of first sampling and then compression. Data compression is realized at the same time of sampling, so that only a small amount of data is sampled, which greatly frees the sampling devices and memory devices. With the development of communication technology, broadband communication and high-frequency applications have brought great challenges to sampling equipment, so this paper mainly studies the application of compression sensing in communication. This paper first studies modulation pattern recognition based on sparse representation of signals. Because the traditional modulation recognition algorithm based on wavelet transform needs symbol synchronization and longer data, its computation is large and time-consuming. It is difficult to meet the real-time requirements of battlefield communication. In this paper, wavelet transform is used to extract the information of signal mutation point, according to the different characteristics of sparse wavelet coefficients before and after normalization. The sparsity of signal wavelet transform is used as the feature parameter of classification and recognition, and the process of classification decision is optimized. The algorithm needs no symbol synchronization, the complexity of the algorithm is low and the signal-to-noise ratio (SNR) is low. This algorithm has a higher recognition rate than the traditional modulation recognition algorithm based on wavelet transform. Then we study the broadband spectrum sensing method based on compressed sensing. In the current spectrum sensing research based on compressed sensing we usually use the known spectral sparsity as the prior information. The channel sparsity in actual spectrum sensing is unknown and time-varying. In this paper, a sparse adaptive wideband spectrum sensing algorithm is proposed, which uses distributed compression sensing and RIP properties to pre-estimate the sparsity, and obtains the spectrum support set by updating the confidence coefficient. The simulation results show that the proposed algorithm performs better than the spectral sensing algorithm with known sparsity at low SNR. Finally, broadband compressed spectrum sensing based on 1-bit is studied. The node data transmission in distributed spectrum sensing network occupies a certain network communication bandwidth. In this paper, a distributed 1-bit compressed spectrum sensing algorithm is proposed, which quantifies the node-sensing data by 1-bit quantization, only preserves the symbol information, and simplifies the data transmission and storage. The simulation results show that the algorithm is effective.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7

【相似文献】