基于压缩感知的多用户检测技术研究

发布时间：2018-05-14 21:23

本文选题：水下无线传感器网络 + 压缩感知　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：海洋蕴藏着多种多样的丰富资源,是人类维持可持续发展的重要一环。随着人类探索和开发海洋步伐的加快,探索海洋的关键性技术之一的水下无线传感器网络成为研究的热点。由于水下环境复杂多变,恶劣的通信环境决定水下无线传感器网络是一种典型的带宽和能耗受限的网络。压缩感知理论是一种新型的信息采集理论,其利用原始信号的稀疏性,并通过降维观测值,即可实现信号的精确重构。压缩感知理论能够广泛的应用于传感器网络的采集、多址接入、多用户检测等方面,为设计高能耗的传感器网络提供一种新的研究方案。本文主要研究基于压缩感知的水下传感器网络数据收集系统,着重研究基于压缩感知的多用户检测技术,并结合压缩感知在采集和多用户检测的优点构建高能效的数据采集方案。首先,研究零星通信场景下的基于压缩感知多用户检测理论。介绍零星通信场景的概念及应用,总结通用多用户系统模型,并分析零星通信场景下传感器节点数据空域稀疏特性以及压缩感知理论进行节点活跃状态和数据检测理论的可行性及优点。本文针对码分多址CDMA以及交织区分多址IDMA接入方式下的压缩感知多用户检测技术进行研究。在CDMA多址接入方式下,总结CDMA多用户系统模型。重点介绍三个贪婪匹配追踪类算法,分别是正交匹配追踪算法OMP、正交最小二乘算法OLS以及利用原始信号块稀疏特性的群正交匹配追踪算法GOMP。通过仿真实验分析压缩感知算法与传统多用户检测算法的性能对比,压缩感知测量矩阵作为扩频码的性能分析以及三个贪婪匹配追踪算法的多用户检测性能。在IDMA多址接入方式下,总结IDMA多用户系统模型,结合压缩感知算法的活跃状态和传统的逐码片CBC多用户检测算法的数据检测的优点提出一种CS-CBC算法。仿真分析IDMA多址接入方式的性能以及验证CS-CBC算法优良的多用户检测性能。最后,针对海洋环境信息收集传感器网络,通过连续利用传感器节点数据在频域和空域上的稀疏性,提出一种基于双域压缩感知的传感器网络高能效数据收集方案DCS。其采用随机采集方式选择部分节点采集数据并传输,汇聚节点利用压缩感知多用户检测算法联合检测节点活跃状态及数据,然后利用成功接收数据包进行信息图重构。DCS数据收集方案结合随机采集及压缩感知多用户检测的高能效收集和传输的优点,仿真分析DCS方案在节省带宽和能耗性能上的优势。
[Abstract]:The ocean contains a variety of rich resources, which is an important part of human sustainable development. With the rapid development of ocean exploration, underwater wireless sensor network (WSNs), one of the key technologies in ocean exploration, has become a hot topic. Because the underwater environment is complex and changeable, the poor communication environment determines that the underwater wireless sensor network is a typical network with limited bandwidth and energy consumption. Compression sensing theory is a new kind of information acquisition theory, which utilizes the sparsity of the original signal and realizes the accurate reconstruction of the signal by reducing the dimension observation value. Compression sensing theory can be widely used in sensor network acquisition, multiple access, multi-user detection and other aspects, which provides a new research scheme for the design of high-energy sensor networks. This paper mainly studies the data collection system of underwater sensor networks based on compressed sensing, focusing on the multi-user detection technology based on compressed sensing. Combining the advantages of compression sensing in acquisition and multi-user detection, a high energy efficiency data acquisition scheme is constructed. Firstly, the theory of compressing-aware multi-user detection in sporadic communication scenarios is studied. This paper introduces the concept and application of sporadic communication scene, summarizes the general multiuser system model, The spatial sparsity characteristics of sensor nodes and the feasibility and advantages of compression sensing theory for node active state and data detection theory in sporadic communication scenarios are analyzed. In this paper, the compressed sensing multiuser detection technology based on CDMA CDMA and interleaved differentiated multiple access (IDMA) is studied. Under the CDMA multiple access mode, the CDMA multi-user system model is summarized. This paper mainly introduces three greedy matching tracking algorithms: OMP orthogonal matching tracking algorithm OLS and GOMP-based group orthogonal matching tracing algorithm based on the sparse property of original signal blocks. The performance of compressed sensing algorithm and traditional multi-user detection algorithm is compared by simulation experiments. The performance of compressed sensing measurement matrix as spread spectrum code and the performance of three greedy matching tracking algorithms are analyzed. This paper summarizes the model of IDMA multiuser system under IDMA multiple access mode, and proposes a CS-CBC algorithm combining the active state of compression sensing algorithm and the advantages of traditional CBC multiuser detection algorithm. Simulation analysis of the performance of IDMA multiple access and verification of the performance of CS-CBC algorithm for multiuser detection. Finally, for marine environmental information collection sensor networks, by continuously utilizing the sparsity of sensor node data in frequency domain and spatial domain, a high energy efficiency data collection scheme for sensor networks based on two-domain compression sensing is proposed. It selects some nodes to collect data and transmit them by random acquisition. The convergent node uses compression sensing multi-user detection algorithm to jointly detect the active state and data of nodes. Then, the advantages of DCS scheme in bandwidth saving and energy consumption performance are analyzed by combining the advantages of random collection and compression sensing multi-user detection, high energy efficiency collection and transmission, using successfully received data packets to reconstruct the information graph.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.3;P715.5

【相似文献】