Volterra卫星信道盲均衡算法

发布时间：2018-04-04 00:03

  本文选题：卫星信道　切入点：盲均衡　出处：《南京信息工程大学》2016年硕士论文

【摘要】：在数字通信系统中不可避免地存在非线性效应,这种非线性将产生严重的幅度畸变和码间干扰,对通信系统数据传输速率的提高是一个阻碍,因此在接收端需要通过非线性均衡以克服码间干扰和幅度畸变。这种非线性效应的产生主要来源于卫星内部放大器的非线性特性,本论文的主要内容是对非线性卫星信道均衡做了一个系统的研究,针对非线性卫星信道产生的干扰和畸变,在传统盲均衡算法的基础上,结合Volterra级数,MMSE,Turbo,模糊神经网络,复数神经网络等一系列方法,提出了多种基于非线性Volterra卫星信道的盲均衡算法,并通过理论分析和计算机仿真实验证明所提算法的有效性。具体研究内容如下：1.针对卫星信道的非线性,用Volterra模型模拟非线性卫星信道,利用逆滤波原理并结合盲均衡算法分析了传统Volterra均衡器并研究了改进的Volterra均衡器。理论分析和仿真结果表明,改进的Volterra均衡器的运算量有所减小。2.非线性产生的码间干扰是影响卫星信道通信的重要因素之一,运用Volterra级数分解来表示非线性信道,为了能够同时消除线性和非线性干扰,推导了基于MMSE的Turbo均衡算法以及无先验信息和低复杂度的两种近似算法；而且为了提高带宽利用率,引入盲均衡算法,提出了基于线性MMSE的迭代Turbo盲均衡算法。仿真结果表明,基于MMSE的迭代Turbo盲均衡算法误码性能有明显提高。3.针对传统常模算法收敛速度与剩余均方误差之间的矛盾及传统神经网络参数太多、复杂度高的问题,提出了基于模糊神经网络控制的复数神经多项式常模盲均衡算法。该算法中的复数神经多项式模块包含单层神经网络和非线性处理器,结构简单,比传统的多层神经网络或递归神经网络参数少复杂度低；而且,利用模糊神经网络模块设计的模糊规则控制迭代步长,提高了步长控制的精度。理论分析和仿真结果表明,该算法具有简单的系统结构、较快的收敛速度和较小的稳态误差,较好的解决了因为参数多而造成的高复杂度问题,而且克服了收敛速度与均方误差之间的矛盾。
[Abstract]:The nonlinear effect inevitably exists in the digital communication system, which will produce serious amplitude distortion and inter-symbol interference, which is a hindrance to the improvement of the data transmission rate of the communication system.Therefore, nonlinear equalization is needed at the receiver to overcome inter-symbol interference and amplitude distortion.This nonlinear effect mainly comes from the nonlinear characteristics of the satellite internal amplifier. The main content of this paper is to do a systematic study on the nonlinear satellite channel equalization, aiming at the interference and distortion caused by the nonlinear satellite channel.On the basis of the traditional blind equalization algorithm and a series of methods such as Volterra series MMSE Turbo, fuzzy neural network and complex neural network, several blind equalization algorithms based on nonlinear Volterra satellite channel are proposed.The effectiveness of the proposed algorithm is proved by theoretical analysis and computer simulation.The specific contents of the study are as follows: 1.Aiming at the nonlinearity of satellite channel, the nonlinear satellite channel is simulated by Volterra model. The traditional Volterra equalizer is analyzed by using the inverse filtering principle and the blind equalization algorithm, and the improved Volterra equalizer is studied.Theoretical analysis and simulation results show that the computational complexity of the improved Volterra equalizer is reduced by .2.The nonlinear inter-symbol interference (ISI) is one of the most important factors affecting satellite channel communication. The Volterra series decomposition is used to represent the nonlinear channel in order to eliminate both linear and nonlinear interference at the same time.The Turbo equalization algorithm based on MMSE and two approximate algorithms without prior information and low complexity are derived, and in order to improve the bandwidth utilization, a blind equalization algorithm is introduced, and an iterative Turbo blind equalization algorithm based on linear MMSE is proposed.Simulation results show that the BER performance of iterative Turbo blind equalization algorithm based on MMSE is significantly improved.Aiming at the contradiction between convergence speed and residual mean square error of traditional constant norm algorithm and the problem that the traditional neural network has too many parameters and high complexity, a complex neural polynomial constant modulus blind equalization algorithm based on fuzzy neural network control is proposed.The complex neural polynomial module in this algorithm consists of a single layer neural network and a nonlinear processor, which has a simple structure and less complexity than the traditional multilayer neural network or recurrent neural network.The iterative step size of fuzzy rule control is designed by using fuzzy neural network module, and the precision of step size control is improved.Theoretical analysis and simulation results show that the algorithm has simple system structure, faster convergence speed and smaller steady-state error, and solves the problem of high complexity caused by many parameters.Moreover, the contradiction between convergence rate and mean square error is overcome.
【学位授予单位】：南京信息工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.5;TN927.2
，

本文编号：1707491