模型与数据结合的浅海时变水声信道估计与均衡

发布时间：2018-05-03 17:25

本文选题：水声通信 + 信道估计　；参考：《浙江大学》2014年博士论文

【摘要】：水声通信是水下环境中实现无线传输的主要手段。水中声传播受三个主要因素制约：随信号频率增大的衰减,时变的传播多径以及较慢的传播速度。水声传播的物理特性造成水声信道,特别是浅海声信道可用带宽窄,时延和多普勒扩展较大且随环境变化而变化,使得水声通信面临许多陆地无线电通信中不曾遇到的技术难点。论文针对浅海时变水声信道下的可靠通信问题,开展水声通信信道估计-信道均衡技术的联合研究。论文挖掘信道的稀疏特性,采用压缩传感算法估计稀疏信道,提高信道估计性能的同时降低算法的复杂度。引入时反处理和频域上与之对偶的频反处理简化信道均衡过程。此外,论文将对信道估计和均衡问题的研究由单入单出(Single-Input Single-Output-SISO)通信系统扩展到多入多出(Multiple-Input Multiple-Output-MIMO)通信系统。在理论算法研究的基础上,还进行了通信实验系统设计并在多次湖海试验中得到使用；进而通过实验数据处理分析,验证了论文方案的有效性及实际性能。信道特性的研究是分析和设计水声通信系统的基础。论文针对浅海水声通信信道时延多普勒双扩展的特性对信道进行建模。首先利用水声信道的稀疏特性,用几条主要的分离路径近似表示信道,然后用多项式分别拟合各到达路径的幅度时变和时延时变,并以这些参数化表示为基础,推导出描述通信信道输入输出关系的数学模型。与之前描述时延多普勒双扩展特性的信道模型相比,本论文推导的路径时变参数化模型考虑了路径幅度的时变及由时延变化引起的成型脉冲上采样点的偏移这些被其它模型忽视的问题,能够更好地近似实际信道的时变特性。路径时变参数化模型用有限的参数描述通信信道输入输出关系,将高维的信道估计问题转化为有限维度的信道参数估计问题,并可采用压缩传感算法求解。注意到重采样和载波频移补偿后的等效信道时变慢的特点,论文提出了一种分步估计时延和多普勒的信道估计方法,该方法使参数的搜索范围从二维时延-多普勒平面退化到一维时延／多普勒区间,减少了字典的维数,从而可降低计算复杂度和存储空间需求。同时,相比于时延和多普勒同步估计方法,将时延和多普勒分开估计又有效避免了参数间的耦合,可以得到更精确的信道估计结果。仿真和实验数据处理结果验证了时延多普勒分步估计方法的性能。浅海水声信道中时延扩展引起严重的码间干扰(Inter-Symbol Interference-ISI),是限制浅海水声通信性能的重要原因之一。针对这一问题,论文研究了水声通信中的信道均衡技术。时间反转处理利用环境本身压缩时域上扩展的信号,是一种利用环境物理特性实现通信信号处理的方法。时反处理并无法完全消除ISI,但是将时反处理和信道均衡结合可以在很大程度上简化均衡器的设计。论文将基于路径时变参数化模型的时变信道响应估计应用于时反处理,把基于信道时不变假设的传统时反通信推广到时变信道。此外,论文还提出了频域上与时间反转对偶的频率反转概念。频反处理可以压缩多普勒效应引起的信号频率扩展以消除正交频分复用(Orthogonal Frequency Division Multiplexing-OFDM)通信中的载波间干扰(Inter-Carrier Interference-ICI)。论文进一步研究了MIMO水声通信系统中的信道估计与均衡问题。论文将路径时变参数化模型扩展到MIMO信道,然后由此模型推导出离散时间MIMO信道输入输出关系,并将时延多普勒分步估计方法推广到对MIMO水声通信信道的估计中。在对MIMO水声通信信道均衡的研究中,论文将MIMO体制与时反技术结合,利用时反的空间聚焦特性实现与多个用户的同时通信;将SISO水声通信系统中的判决反馈均衡器(Decision Feedback Equalizer-DFE)扩展到MIMO系统中,在反馈部分引入其它路的发射符号在消除ISI的同时消除信道间干扰(Cochannel Interference-CCI)。论文还将频反频域均衡方法扩展到MIMO通信中,与时反处理一样,频反处理也能达到空间聚焦的效果,消除CCI,是MIMO-OFDM中一种潜在有效的频域均衡手段。
[Abstract]:Underwater acoustic communication is the main means to realize wireless transmission in underwater environment. Acoustic propagation in water is restricted by three main factors: attenuation with signal frequency, time-varying propagation multipath and slower propagation speed. The physical characteristics of underwater acoustic propagation cause underwater acoustic channel, especially in shallow sea sound channel with narrow bandwidth, delay and Doppler expansion. Large and changing with the environment, underwater acoustic communication is faced with many technical difficulties that have not been encountered in terrestrial radio communications.
In this paper, a joint study of channel estimation - channel equalization in underwater acoustic channel is carried out in light of the time-varying underwater acoustic channel. The paper excavate the sparse characteristics of the channel and estimate the sparse channel using the compression sensing algorithm to improve the performance of the channel estimation while reducing the complexity of the algorithm. The dual frequency inverse processing simplifies the channel equalization process. In addition, this paper extends the study of channel estimation and equalization from single out single out (Single-Input Single-Output-SISO) communication system to multiple entry and multiple (Multiple-Input Multiple-Output-MIMO) communication system. Based on the study of theoretical calculation, the communication experiment system is also set up. It has been used in many experiments of lake and sea. The validity and practical performance of the scheme are verified by experimental data processing and analysis.
The research of channel characteristics is the basis of analysis and design of underwater acoustic communication system. In this paper, the channel is modeled by the characteristics of channel delay Doppler double extension in shallow sea acoustic communication channel. Firstly, the sparse characteristic of underwater acoustic channel is used to approximate the channel with several main separation paths, and then the amplitude of each arrival path is fitted by polynomial respectively. On the basis of these parameterized representations, the mathematical model describing the input and output relations of communication channels is derived. Compared with the channel model describing the doubly extended characteristic of time delay Doppler, the time-varying parameterized model derived from this paper takes into account the time variation of the path amplitude and the formation caused by the change of time delay. The offset of the upper sampling points of the pulse, which is ignored by other models, can better approximate the time-varying characteristics of the actual channel.
The path time-varying parameterized model describes the input and output relation of communication channel with limited parameters, and transforms the high dimension channel estimation into the finite dimension channel parameter estimation problem, and can use the compression sensing algorithm to solve the problem. Step estimation time delay and Doppler's channel estimation method, which reduces the search range from two-dimensional delay to Doppler plane to one dimension time delay / Doppler interval, reduces the dimension of the dictionary and reduces the computational complexity and storage space. At the same time, the time delay and the Doppler synchronization method are compared. The split estimation can effectively avoid the coupling between parameters and obtain more accurate channel estimation results. The simulation and experimental data processing results verify the performance of the time-delay Doppler step estimation method.
Time delay expansion in shallow sea underwater acoustic channel causes serious inter code interference (Inter-Symbol Interference-ISI), which is one of the important reasons for limiting the performance of shallow sea acoustic communication. The communication signal processing method is realized by the environmental physical properties. The time inverse processing can not completely eliminate the ISI, but the combination of time reversal and channel equalization can simplify the design of equalizer to a large extent. In addition, the frequency reversal concept of frequency reversal in frequency domain with time reversal is also proposed. Frequency inverse processing can compress the signal frequency expansion caused by the Doppler effect to eliminate intercarrier interference in Orthogonal Frequency Division Multiplexing-OFDM communication (Int Er-Carrier Interference-ICI).
The paper further studies the channel estimation and equalization in MIMO underwater acoustic communication system. This paper extends the path time-varying parameterized model to the MIMO channel, and then derives the input and output relation of the discrete time MIMO channel, and extends the time-delay Doppler step estimation method to the estimation of the MIMO underwater communication channel. In the MIM In the study of channel equalization for O underwater acoustic communication, the paper combines the MIMO system with the time inverse technology and uses the time inverse spatial focusing characteristics to realize the simultaneous communication with multiple users. The decision feedback equalizer (Decision Feedback Equalizer-DFE) in the SISO underwater acoustic communication system is extended to the MIMO system, and the other channel emitter is introduced in the feedback part. ISI eliminates inter channel interference (Cochannel Interference-CCI) at the same time. This paper also extends the frequency inverse frequency domain equalization method to MIMO communication. As with time reversal processing, frequency inverse processing can also achieve the effect of space focusing and eliminate CCI. It is a potentially effective frequency domain equalization method in MIMO-OFDM.

【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.3

【参考文献】