卫星信道盲均衡算法研究

发布时间：2018-05-30 02:14

本文选题：盲均衡 + 循环平稳　；参考：《南京大学》2014年硕士论文

【摘要】：卫星通信是目前最有应用前途的通信技术之一,该技术通信距离远,对地面上的各种电波干扰很不敏感,普遍应用于军事通信、卫星电视广播等方面。然而卫星信道的群时延特性在很大程度上影响了数据传输质量,使得码间串扰非常严重。卫星通信系统的广泛应用和迅速发展,使得人们对其均衡性能的要求不断提高。均衡技术能够有效地改善因信道的各种噪声和干扰引起的码元失真,在知晓信道参数的情况下,可以利用训练序列,通过比较已知的发射序列和接收到的序列的误差来迭代调整均衡器的抽头系数,使得均衡器经过多次调节最终可以逼近理想状态。但是大量的发送训练序列会浪费较多带宽,引起传输有效信息效率降低。在这种情况下,能够不需要知道信道参数不需要发送训练序列,仅利用信号的一些先验信息就能实现信道均衡的盲均衡技术越来越成为人们关注的重点。本文针对卫星信道的群时延传输特性,重点研究了卫星信道环境中的盲均衡技术,所做工作有以下几点：首先,本文阐述了卫星信道的传输特性,点明了卫星传输对均衡技术要求的迫切性,并根据卫星信道的传输特性选取了具有类似特性的信道参数作为以后仿真的适用信道。其次本文介绍了均衡器的基本知识,包括传统自适应均衡技术和盲均衡技术的有关概念,相应结构及其相互区别之处。介绍了最基本的均衡算法以及衡量均衡算法性能的相关指标等内容,为下文设计分析新型盲均衡算法提供理论基础。再次针对不同类型的盲均衡算法,本文概要介绍了基于二阶循环平稳统计量的盲均衡算法。首先介绍了循环平稳信号的基础知识,阐述了一阶循环平稳信号和二阶循环平稳信号的相关特性,并根据循环平稳信号的特性分析了基于调制引入循环平稳的单信道盲辨识法,这种盲辨识法无需知晓信号参数,仅仅利用发送信号时引入循环平稳,就可以根据循环平稳特性采用一些数学方法分析并计算出信道参数。针对盲辨识算法所采用的基于联合调制和过抽样引入循环平稳的信号模型,设计了循环维纳均衡器,利用了FROST算法解决了均衡器设计过程中的受限自适应最优化问题。接着本文又介绍了另一类被广泛应用的盲均衡算法Bussgang类盲均衡算法,简单的阐述了Bussgang过程和几种经典的Bussgang类盲均衡算法,重点讨论了恒模盲均衡算法,针对恒模盲均衡算法对相位旋转调整不足的缺点引出了修正恒模盲均衡算法,对修正恒模算法采用固定步长这个缺点提出改进,利用一个于剩余误差有关的非线性函数来控制修正恒模算法的迭代步长,从而形成了可变步长的修正恒模算法。新算法与修正恒模算法相比收敛速度和收敛精度有了一定的提高。最后,在现有的单模算法的基础上引入了多模盲均衡算法。介绍了多模算法的构成原理和优缺点。在多模算法中最普遍应用的双模盲均衡算法的基础上,提出了一种新型双模盲均衡算法。新算法同时利用切换模式和加权模式将可变步长修正恒模算法和判决引导最小均方算法相联合,与单模算法相比,新双模算法有着更好的灵活适用性和更高的均衡性能。
[Abstract]:Satellite communication is one of the most promising communication technologies. This technology has a long communication distance and is very insensitive to various interference on the ground. It is widely used in military communications, satellite TV broadcasting and so on. However, the delay of the group of satellite channels greatly affects the quality of data transmission and makes inter code crosstalk very strict. With the wide application and rapid development of satellite communication system, the demand for the equalization performance is improved continuously. The equalization technology can effectively improve the distortion of the symbol caused by the noise and interference of the channel. In the case of knowing the channel parameters, the training sequence can be used, and the known transmission sequence and reception can be received. The error of the sequence iteratively adjusts the tap coefficient of the equalizer so that the equalizer can eventually approach the ideal state after multiple adjustments, but a large number of training sequences will waste more bandwidth and reduce the efficiency of transmission information. In this case, the training sequence can not be needed without knowing the channel parameters, only The blind equalization technology of channel equalization using some prior information of the signal has become the focus of attention. This paper focuses on the blind equalization technology in satellite channel environment for the group delay transmission characteristics of satellite channels, and the following work has been done.
First, this paper describes the transmission characteristics of satellite channel, points out the urgency of satellite transmission for equalization technology, and selects channel parameters with similar characteristics according to the transmission characteristics of satellite channel as the suitable channel for future simulation.
Secondly, this paper introduces the basic knowledge of equalizer, including the concepts of the traditional adaptive equalization technology and the blind equalization technology, the corresponding structure and the differences between them, and introduces the basic equilibrium algorithm and the related indexes to measure the performance of the equalization algorithm, which provides the theoretical basis for the design and analysis of the new blind equalization algorithm below.
The blind equalization algorithm based on the two order cyclostationary statistics is introduced in this paper. The basic knowledge of the cyclostationary signal is introduced, and the correlation characteristics of the first order cyclostationary signal and the two order cyclostationary signal are described, and the modulation based on the characteristic of the cyclostationary signal is based on the modulation. The blind identification method of cyclic stationary single channel is introduced. This blind identification method does not need to know the signal parameters. It can be used to analyze and calculate the channel parameters by using some mathematical methods based on the cyclostationary characteristics. A steady signal model is designed, and a circulant Wiener equalizer is designed. The FROST algorithm is used to solve the constrained adaptive optimization problem in the design of the equalizer.
Then this paper introduces another kind of blind equalization algorithm Bussgang blind equalization algorithm, which is widely used. It simply describes Bussgang process and several classical Bussgang blind equalization algorithms, and focuses on the constant modulus blind equalization algorithm, which leads to the correction of the constant modulus blind equalization for the shortcoming of the constant modulus blind equalization algorithm to adjust the phase rotation. A modified constant modulus algorithm is proposed by using a nonlinear function related to the residual error to control the iteration step of the modified constant modulus algorithm, thus forming a modified constant modulus algorithm with variable step length. The new algorithm is compared with the modified constant modulus algorithm. High.
Finally, the multimode blind equalization algorithm is introduced on the basis of the existing single mode algorithm. The principle and the advantages and disadvantages of the multimode algorithm are introduced. A new double mode blind equalization algorithm is proposed on the basis of the most commonly used double mode blind equalization algorithm in the multimode algorithm. The new algorithm simultaneously uses the switching mode and the weighted mode to change the variable step length. The modified constant modulus algorithm is combined with the decision - guided least mean square algorithm. Compared with the single mode algorithm, the new dual mode algorithm has better flexibility and higher equilibrium performance.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.5;TN927.2

【参考文献】