【作者】 李霞；

【导师】 桑恩方；

【作者基本信息】 哈尔滨工程大学 ， 信号与信息处理， 2004， 博士

【摘要】 随着人们开发和利用海洋步伐的加快，水下数字通信技术的研究越来越受到人们的重视，逐渐成为水声研究领域中最具挑战性的课题之一。在带宽严重受限的水声信道中实现高数据率通信，必须使用高带宽利用率调制技术，并结合自适应均衡等技术，以有效地克服信道多途传播产生的符号间干扰。 本文以克服多途干扰、实现高可靠性的水声通信为出发点，以空间分集均衡技术为主要研究内容，所做的工作和贡献主要概括为如下几个方面： ·对典型的线性和非线性自适应均衡器结构和算法进行研究，通过仿真对其性能进行了比较；在此基础上对相应的派生算法进行了仿真和分析。 ·对块均衡、稀疏均衡及基于信道估计的均衡算法进行了阐述，并对几种自适应算法进行了仿真比较，从而挑选出运算量较低的均衡算法。 ·对信号进行分集处理的几种常用技术进行了定量描述，并对空间分集均衡处理中的联合均衡组合器和MMSE均衡组合器两种结构的工作原理进行了详细阐述，证明了后者在计算复杂度方面优于前者。 ·给出了线性分集均衡器和判决反馈分集均衡器两种结构的工作原理，并从理论上证明了后者的最小均方误差性能优于线性空间分集组合器；给出了多通道自适应算法，并对不同算法的性能进行了比较，通过仿真证明了空间分集均衡器与常规均衡器的性能差异。 ·在空间分集均衡的基础上结合了人工神经网络的处理技术，提出了一种可用于空间分集的递归神经网络空间分集均衡器结构；在新采用的BP神经元网络算法中，提出了由神经元输入信号的强度来控制输出层权值的方法，从而增强了信号成分的贡献，使系统的输出信噪比进一步得到提高； ·为了更有效地跟踪水声信道的多普勒频移，提出了将二阶数字锁相环与空间分集均衡相结合的算法，仿真结果表明，采用二阶数字锁相环的空间分集均衡算法可有效补偿由于水声信道时变而引入的载波相位快速变化。 ·为了提高水声信道的利用率，对多用户的MIMO系统进行了理论研究，并对其性能进行了计算机仿真。 ·在吉林松花湖进行了水声通信试验，采用提出改进的算法对湖试数据进行哈尔滨工程大学博士学位论文处理，并与几种常规均衡结构的性能进行了比较，验证了所提出的算法在收敛速度和容错性能方面优于常规空间分集均衡技术。该实验是国内首次进行的采用空间分集技术的水声系统的湖上试验，并获得了良好的试验结果;关键词:水声通信;多途;自适应均衡;空间分集;人工神经网络更多还原

【Abstract】 Nowadays, digital communication through band-limited underwater acoustic channel (UWAC) is of general interest in the field of underwater communication. In order to achieve high data rates on the severely band-limited underwater acoustic channels, bandwidth-efficient modulation techniques must be employed, together with some sophisticated signal processing algorithms, such as adaptive equalization technique for the intersymbol interference caused by channel multipath propagation.To attain reliable data transmission over UWA channels, spatial diversity technology and adaptive equalization technology are selected as major research issue. The main content in this paper are as follows:All kinds of linear and nonlinear adaptive equalization and typical algorithms are studied elaborately. And their performances are compared by simulation. On the base of above algorithms, corresponding derive algorithms are simulated and analyzed. To decrease operation of the equalization algorithm, the block equalization, sparse algorithm and the algorithm based on channel estimated are deduced in detail, also these algorithms simulation and comparison have been done. Several general methods of signal diversity process are described, and the principle of the joint equalization combiner and MMSE equalization combiner are deduced in detail to test computation complexity of the structure. The principles of the linearity diversity equalizer and decision feedback diversity equalizer are presented, and the MMSE of the latter is proved to be better than the former. Multi-channel adaptive algorithms are presented and comparison, and the performance difference between the spatial diversity equalizer and general equalizer are testified by simulation. On the base of spatial diversity equalization the artificial neural network technique is combined. The recursion neural network diversity equalizer ispresented, also the corresponding improved adaptive update algorithm, and the validity of the structure and algorithm are testified. To track the Doppler shift, the second order digital phase lock loop is combined with the spatial diversity equalizer. The simulation and test results show that the method can compensate the fast change of the carrier wave phase caused by the time-vary of UWAC.To improve the efficiency of UWAC, the multi input and multi output (MEMO) system is studied, and the computer simulation is done for the performance. The acoustic communication test is completed in Songhua Lake in Jilin province. Using these data, we demonstrate the efficiency of several equalization algorithms, including compassion study and evaluation of their performance.更多还原