基于自适应双链DNA遗传优化的盲均衡算法

发布时间：2018-10-11 10:19

【摘要】：在无线通信中,由于通信信道复杂多变而导致的失真和有限带宽所引起的码间干扰(Inter-symbol Interference, ISI)严重影响通信质量,因此在通信系统的接收端需要适当的补偿这些因素所带来的影响。由于盲均衡技术不需要发送训练序列,节省了带宽,因此可以有效地克服码间干扰和提高通信速度。在信号处理领域中,盲均衡技术已经成为一个研究热点。本文主要针对传统盲均衡算法收敛速度慢、稳态误差大的缺陷,利用DNA遗传算法和正交小波变换等手段,对盲均衡算法的均衡性能进行优化,研究内容主要包括以下几个方面：(1)提出了基于DNA遗传优化的常模盲均衡算法。传统盲均衡算法采用最速下降法对均衡器权向量进行更新,然而最速下降法要求代价函数必须满足连续、可导的条件,并且容易陷入局部极值,从而导致盲均衡算法收敛速度慢,稳态误差大。因此,针对这些缺点,利用DNA遗传算法的全局搜索能力对均衡器的权向量进行优化,避免常模盲均衡算法出现局部收敛,改善了盲均衡算法的均衡性能。仿真结果表明了该算法的有效性。(2)提出了基于禁忌搜索的自适应双链DNA遗传优化常模盲均衡算法。由于在一般的DNA遗传算法中,种群的交叉操作的概率都是定值,不能随种群变化而做适当的调整,因此传统的DNA遗传算法搜索效率不高。另外,DNA遗传算法在对问题最优解进行搜索时,有时会出现重复搜索,同样影响了算法的搜索效率。因此针对这些缺陷,结合禁忌搜索算法,同时采用DNA遗传算法中的交叉操作概率随种群进化代数变化的方法,形成了基于禁忌搜索的自适应双链DNA遗传优化常模盲均衡算法。(3)提出了基于禁忌搜索的自适应双链DNA遗传优化多模盲均衡算法。针对传统的多模盲均衡算法对高阶多模调制信号均衡效果不好的问题,采用正交小波变换,将基于禁忌搜索的自适应双链DNA遗传算法应用到多模盲均衡算法中,从而形成了基于禁忌搜索的自适应双链DNA遗传优化多模盲均衡算法。由于基于禁忌搜索的自适应双链DNA遗传算法具有更快的搜索效率和更好局部搜索特性,因此,将DNA遗传算法应用到多模盲均衡算法中,能够显著地提高多模盲均衡算法的性能。(4)提出了基于多种群进化的禁忌搜索DNA遗传优化小波分数间隔多模盲均衡算法。为了进一步提高DNA遗传算法的搜索效率和小波分数间隔多模盲均衡算法的均衡效果,利用禁忌搜索算法和多种群进化策略,对DNA遗传算法进行改进,并且将改进后的算法应用到小波分数间隔多模盲均衡算法中,从而形成了基于多种群进化的禁忌搜索DNA遗传优化小波分数间隔多模盲均衡算法。该算法将DNA种群分为多个子种群,每个子种群侧重于不同的搜索目的,即主种群侧重于局部搜索,辅助种群侧重于全局搜索,并且在三个种群中分别采用禁忌交叉操作和不同的变异操作,提高了DNA遗传算法的搜索能力,从而改善了小波分数间隔多模盲均衡算法的性能。
[Abstract]:In wireless communication, inter-symbol interference (ISI) due to the complexity of communication channel and the inter-symbol interference (ISI) caused by the limited bandwidth seriously affect the communication quality, and therefore, it is necessary to compensate the influence of these factors at the receiving end of the communication system. Since the blind equalization technique does not need to send a training sequence, the bandwidth is saved, so that inter-code interference and communication speed can be effectively overcome. Blind equalization has become a hot topic in the field of signal processing. Aiming at the defects of slow convergence speed and large steady-state error of traditional blind equalization algorithm, the equalization performance of blind equalization algorithm is optimized by means of DNA genetic algorithm and orthogonal wavelet transform. (1) A norm blind equalization algorithm based on DNA genetic optimization is proposed. The traditional blind equalization algorithm is used to update the equalizer weight vector by steepest descent method. However, the steepest descent method requires that the cost function must satisfy continuous and guide conditions, and can easily fall into local extreme value, leading to the slow convergence speed and large steady-state error of the blind equalization algorithm. Therefore, aiming at these disadvantages, the weight vector of the equalizer is optimized by using the global search capability of the DNA genetic algorithm so as to avoid local convergence of the norm blind equalization algorithm and improve the equalization performance of the blind equalization algorithm. The simulation results show the validity of the algorithm. (2) An adaptive double-stranded DNA genetic optimization norm blind equalization algorithm based on tabu search is proposed. Because of the common DNA genetic algorithm, the probability of cross operation of the population is fixed value, and can not be properly adjusted with the population change, so the traditional DNA genetic algorithm search efficiency is not high. In addition, when searching the optimal solution of the problem, the DNA genetic algorithm sometimes repeats the search, and also influences the search efficiency of the algorithm. Therefore, the adaptive double-stranded DNA genetic optimization norm blind equalization algorithm based on tabu search is formed by combining the tabu search algorithm and the method of crossover operation probability in DNA genetic algorithm with the evolution of population evolution. (3) An adaptive double-stranded DNA genetic optimization multi-mode blind equalization algorithm based on tabu search is proposed. Aiming at the problem that the traditional multi-mode blind equalization algorithm has poor equalization effect on the high-order multi-mode modulation signal, the self-adaptive double-stranded DNA genetic algorithm based on the tabu search is applied to the multi-mode blind equalization algorithm by adopting orthogonal small wave transformation, thereby forming a self-adaptive double-stranded DNA genetic optimization multi-mode blind equalization algorithm based on tabu search. Because the self-adaptive double-stranded DNA genetic algorithm based on tabu search has faster search efficiency and better local search characteristics, the DNA genetic algorithm is applied to the multi-mode blind equalization algorithm, and the performance of the multi-mode blind equalization algorithm can be significantly improved. (4) A multi-mode blind equalization algorithm based on multi-group evolution is proposed to search DNA genetic optimization small-wave fractional interval multi-mode blind equalization. In order to further improve the search efficiency of DNA genetic algorithm and the equalization effect of small-wave fractional-interval multi-mode blind equalization algorithm, using tabu search algorithm and multi-group evolution strategy, the DNA genetic algorithm is improved. and applying the improved algorithm to the small-wave fractional-interval multi-mode blind equalization algorithm, thereby forming a tabu search DNA genetic optimization small wave fractional interval multi-mode blind equalization algorithm based on a plurality of group evolution. The algorithm divides the DNA population into a plurality of sub-populations, each sub-population focuses on different search purposes, namely, the main population is focused on local search, the auxiliary population is focused on global search, and taboos cross operation and different mutation operations are respectively adopted in the three populations, improves the search capability of the DNA genetic algorithm, thereby improving the performance of the small-wave fractional-interval multi-mode blind equalization algorithm.
【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】：2015
【分类号】：TN911.5

【参考文献】