水声通信中信道编码的研究与DSP实现

发布时间：2018-10-20 09:18

【摘要】：人们开发和利用海洋资源的一切活动,都与水声通信技术有着紧密的联系。经过不断地发展,水声通信已经形成了完整的系统体系。但由于水声信道具有可用频带窄、强多途、高噪声的特点,所以未编码的水下通信系统误码率通常较高,很难满足实际需求。因此信道编码技术的研究对实现高质量的水声通信具有十分重要的意义。信道编码通过在信息序列中加入校验序列,使接收端能够对接收序列进行纠错,从而减少由信道干扰带来的错误,进一步降低了水声通信的误码率。本文主要针对信道编码中的RS码、卷积码和Turbo码进行Matlab仿真研究和DSP实现。论文的第一个主要部分是信道编码的Matlab仿真。首先是对RS码的仿真,主要是通过改变编码长度和纠错能力两个参数对其误码率进行分析;接着是对卷积码的仿真,分析了码率、生成矩阵、约束长度、回溯长度、判决方式以及软判决时的量化值Q对卷积码误码率的影响;在对Turbo码进行仿真时,不仅对码率、分量码、迭代次数、交织长度四个参数进行了分析,还同时对Turbo码的Log_MAP、Max_Log_MAP和SOVA三种译码算法进行比较,了解不同译码算法对性能的影响;最后将三种信道编码在高斯白噪声信道下对比,并加入水声信道冲激响应进行仿真,分析水声信道下不同信道编码的性能特点。论文的另一个主要部分是三种信道编码在DSP上的实现,算法实现的硬件平台选用的是TMS320C6455型DSP芯片。首先对RS码、卷积码和Turbo码编解码的软件实现流程进行了详细的介绍,主要通过程序流程图的形式给出,同时对编解码过程中变量的存储空间进行了分配。然后向编码后的数据中加入一些错误,通过解码后的数据与编码前数据进行比较,验证三种编码成功在DSP上实现。最后测出三种信道编码的编码耗时和解码耗时,对算法复杂度进行比较。通过仿真和DSP实现的结果可以看出,在纠错性能方面,Turbo码的性能要比RS码、卷积码好很多;但在实现的复杂度方面,Turbo码要比RS码、卷积码更加复杂。在水声通信系统中,为了提高信息传递的可靠性,满足高速的数据通信和高效的资源利用,选用Turbo码的效果要更好一些。
[Abstract]:All activities of developing and utilizing marine resources are closely related to underwater acoustic communication technology. After continuous development, underwater acoustic communication has formed a complete system. However, the underwater acoustic channel has the characteristics of narrow available band, strong multi-path and high noise, so the uncoded underwater communication system usually has a high bit error rate, which is difficult to meet the actual needs. Therefore, the research of channel coding technology is very important to the realization of high quality underwater acoustic communication. Channel coding can correct the error of the received sequence by adding a check sequence to the information sequence, thus reducing the error caused by channel interference and further reducing the bit error rate of underwater acoustic communication. In this paper, the RS code, convolutional code and Turbo code in channel coding are studied by Matlab simulation and DSP implementation. The first part of this paper is the Matlab simulation of channel coding. The first part is the simulation of RS code, which mainly analyzes the bit error rate by changing the coding length and error correcting ability, and then analyzes the code rate, the generation matrix, the constraint length and the backtracking length of the convolutional code. In the simulation of Turbo code, not only four parameters, such as bit rate, component code, iteration number and interleaving length, are analyzed, but also the error rate of convolutional code is analyzed. At the same time, the Log_MAP,Max_Log_MAP and SOVA decoding algorithms of Turbo codes are compared to understand the influence of different decoding algorithms on the performance. Finally, the three channel codes are compared in Gao Si white noise channel, and the impulse response of underwater acoustic channel is simulated. The performance characteristics of different channel codes in underwater acoustic channels are analyzed. Another main part of this paper is the implementation of three channel codes on DSP. The hardware platform of the algorithm is TMS320C6455 DSP chip. Firstly, the software implementation flow of RS code, convolutional code and Turbo code is introduced in detail, mainly in the form of program flow chart. At the same time, the storage space of variables in the process of coding and decoding is allocated. Then some errors are added to the encoded data, and by comparing the decoded data with the pre-coding data, it is verified that the three codes are successfully implemented on DSP. Finally, the coding time and decoding time of the three channel codes are measured, and the complexity of the algorithm is compared. The results of simulation and DSP implementation show that the performance of Turbo code is much better than that of RS code and convolutional code, but the complexity of Turbo code is more complex than that of RS code and convolutional code. In underwater acoustic communication system, in order to improve the reliability of information transmission, to meet the high speed data communication and efficient resource utilization, the effect of selecting Turbo code is better.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.3

【参考文献】