果蠅優(yōu)化小波盲均衡算法研究
本文選題:盲均衡 + 正交小波變換。 參考:《安徽理工大學》2014年碩士論文
【摘要】:目前,國內(nèi)外對果蠅優(yōu)化算法及其應(yīng)用的研究成果較少,也還未發(fā)現(xiàn)有關(guān)于小波變換理論、果蠅優(yōu)化算法和各類盲均衡算法相融合后應(yīng)用于水聲通信領(lǐng)域的研究報道。本文在深入地研究了果蠅優(yōu)化算法的基礎(chǔ)理論知識之后,將新穎的果蠅算法引入盲均衡技術(shù)之中來優(yōu)化均衡器的性能,并且在分析了果蠅優(yōu)化算法具有的優(yōu)勢和劣勢的基礎(chǔ)之上,結(jié)合當前熱點新技術(shù)、新理論-模擬退火思想、小波變換理論和協(xié)同進化策略,對果蠅優(yōu)化算法進行了改進,將改進后的新算法嘗試應(yīng)用于水聲通信盲均衡技術(shù)中以達到進一步優(yōu)化算法的均衡性能來實現(xiàn)提高水聲信號傳輸效率的研究目的。本課題研究具體完成了以下工作: 1.果蠅優(yōu)化的小波盲均衡算法 傳統(tǒng)常數(shù)模盲均衡算法被廣泛應(yīng)用于水聲通信領(lǐng)域,它是通過利用隨機梯度下降的搜索方式來調(diào)整均衡器系數(shù)的,這種搜索方式不僅容易導致CMA陷入局部尋優(yōu),無法準確找到全局極值點而且還對均衡器的代價函數(shù)有連續(xù)、可導的要求。果蠅優(yōu)化算法具有很強的全局搜索能力,將果蠅優(yōu)化算法引入CMA中提出了果蠅優(yōu)化小波盲均衡算法,可避免傳統(tǒng)CMA搜索方法存在的缺陷,正交小波變換可抑制信號之間自相關(guān)性。 2.果蠅優(yōu)化的小波自適應(yīng)軟約束常模盲均衡算法 均衡復(fù)雜的水聲信道時傳統(tǒng)常模盲均衡算法收斂速度很慢、穩(wěn)態(tài)誤差也比較大。自適應(yīng)軟約束常模盲均衡算法的均衡效果優(yōu)于CMA,處理水聲信號時收斂速度明顯快于CMA、穩(wěn)態(tài)誤差也相對較小,但SCS-CMA搜索最優(yōu)權(quán)向量的方式仍和常數(shù)盲均衡相同,都采用的隨機梯度下降法,常易陷入局部收斂。果蠅優(yōu)化小波自適應(yīng)軟約束常模盲均衡算法是在SCS-CMA中融入果蠅優(yōu)化算法和小波變換理論,運用正交小波變換預(yù)處理均衡器的輸入信號來達到去除噪聲和降低輸入信號自相關(guān)性的作用,運用果蠅優(yōu)化算法求解均衡器的代價函數(shù),并用SFOA迭代搜索所得的最優(yōu)權(quán)向量初始化均衡器,該算法的均衡效果明顯優(yōu)于CMA。 3.模擬退火-果蠅混合算法優(yōu)化小波廣義自適應(yīng)多模盲均衡算法 一般來說,我們在盲均衡技術(shù)中運用多模盲均衡算法來處理高階QAM信號,而傳統(tǒng)多模盲均衡算法的均衡效果越來越不能滿足日益增長的實際應(yīng)用的需求。針對MMA和果蠅算法存在的缺點,本文提出了模擬退火-果蠅混合算法優(yōu)化小波廣義自適應(yīng)多模盲均衡算法。這種新的算法結(jié)合模擬退火這一新技術(shù)與果蠅算法兩者的優(yōu)勢,利用局部搜索能力強的模擬退火技術(shù)解決果蠅優(yōu)化算法搜索復(fù)雜的大規(guī)?臻g時易陷入局部收斂的問題。模擬退火-果蠅混合優(yōu)化算法能夠精確快速地找到最優(yōu)權(quán)向量,加快算法的穩(wěn)定收斂速度,降低穩(wěn)態(tài)誤差。使用正交小波對均衡器的每路輸入信號進行分解來除噪去信號的相關(guān)性,進一步改善了廣義離散自適應(yīng)多模盲均衡器的性能,新算法更能有效地均衡高階QAM信號。 4.小波盲均衡多果蠅群協(xié)同優(yōu)化算法 (1)多果蠅群協(xié)同優(yōu)化算法 果蠅優(yōu)化算法的尋優(yōu)精度不高,當尋優(yōu)復(fù)雜搜索區(qū)域時,搜索性能較低,收斂速度較慢。針對果蠅優(yōu)化算法存在的不足,在SFOA中引入?yún)f(xié)同進化思想,提出了多果蠅群協(xié)同優(yōu)化算法。新算法利用并行拓撲的進化結(jié)構(gòu)和正反反饋的信息共享方式來協(xié)同指導整個系統(tǒng)的進化。搜索時,將多個果蠅群作為獨立進化的群體在同時進行搜索中也相互跟蹤對方的全局最優(yōu)解。通過共享對各個果蠅群各自的尋優(yōu)結(jié)果進行評價所得的群體當前最優(yōu)解來指導各個種群在獨立進化的同時協(xié)同進化,直至獲得最優(yōu)解。 (2)多果蠅群協(xié)同優(yōu)化小波常模盲均衡算法 應(yīng)用多果蠅群協(xié)同優(yōu)化算法至盲均衡算法中,在CMA的基礎(chǔ)上融入多蠅協(xié)同的果蠅優(yōu)化算法尋找最優(yōu)權(quán)向量初始化均衡器,正交小波變換理論消噪、減小信號間存在的自相關(guān)性。該算法均衡信號的效果更好。 (3)多果蠅群協(xié)同優(yōu)化小波多模盲均衡算法 針對CMA均衡信號時相位模糊、誤差大、處理高階QAM信號均衡效果差等不足和果蠅優(yōu)化算法所存在的缺陷,分析了可有效糾正信號相位旋轉(zhuǎn)、適用于高階信號均衡的多模盲均衡算法的原理,將其與搜索能力強的多果蠅群協(xié)同優(yōu)化算法和抑制信號相關(guān)性強的小波變換相結(jié)合,提出了一種新算法-多果蠅群協(xié)同優(yōu)化小波多模盲均衡算法。
[Abstract]:At present , the research results of fruit fly optimization algorithm and its application are few , and it has not been found to be related to wavelet transform theory , fruit fly optimization algorithm and various blind equalization algorithms . After studying the basic theory knowledge of fruit fly optimization algorithm , the paper improves the algorithm of fruit fly optimization , and then attempts to apply the improved algorithm to underwater acoustic communication blind equalization technology to achieve the research purpose of improving the transmission efficiency of underwater acoustic signal .
1 . Small - wave blind equalization algorithm for fruit fly optimization
The traditional constant modulus blind equalization algorithm is widely used in the field of underwater acoustic communication .
2 . Optimal Blind Equalization Algorithm for Fruit - fly Optimization with Wavelet Adaptive Soft - constraint
The traditional common mode blind equalization algorithm has a slow convergence speed and a large steady state error when the complex underwater acoustic channel is balanced . The adaptive soft - constrained normal - mode blind equalization algorithm is better than CMA , and the convergence speed of the adaptive soft - constrained normal - mode blind equalization algorithm is obviously faster than CMA , and the steady - state error is relatively small , but the adaptive soft - constrained normal - mode blind equalization algorithm is used in the SCS - CMA to achieve the effect of removing noise and reducing the self - correlation of the input signal .
3 . Simulated annealing - fruit fly hybrid algorithm to optimize the wavelet generalized adaptive multi - mode blind equalization algorithm
In general , we use multi - mode blind equalization algorithm in blind equalization to deal with high - order QAM signals , and the traditional multi - mode blind equalization algorithm is more and more difficult to meet the need of increasing practical application .
4 . Cooperative optimization algorithm for small wave blind equalization multi - fruit fly group
( 1 ) Multi - fruit fly group cooperative optimization algorithm
In the search of complex search area , the search performance is low and the convergence speed is slow . In order to overcome the shortcomings of the optimization algorithm of fruit fly , a cooperative evolutionary algorithm is introduced in SFOA to guide the evolution of the whole system .
( 2 ) Multi - fruit fly group cooperative optimization wavelet normal - mode blind equalization algorithm
In this paper , a multi - fruit fly swarm optimization algorithm is applied to the blind equalization algorithm , and a multi - fly cooperative fruit fly optimization algorithm is integrated on the basis of CMA to find the optimal weight vector initialization equalizer , the quadrature wavelet transform theory denoising and the reduction of the self - correlation between the signals .
( 3 ) Multi - fruit fly group cooperative optimization wavelet multi - mode blind equalization algorithm
This paper analyzes the principle of multi - mode blind equalization algorithm which can effectively correct the phase rotation of the signal , applies to the high - order signal equalization , and combines the multi - fruit fly group cooperative optimization algorithm with the strong search capability and the wavelet transform with strong signal correlation , and proposes a new algorithm - multi - fruit fly group cooperative optimization wavelet multi - mode blind equalization algorithm .
【學位授予單位】:安徽理工大學
【學位級別】:碩士
【學位授予年份】:2014
【分類號】:TN911.5
【參考文獻】
相關(guān)期刊論文 前10條
1 宋煒;劉強;;基于模擬退火算法的過程挖掘研究[J];電子學報;2009年S1期
2 郭業(yè)才;王麗華;;模糊神經(jīng)網(wǎng)絡(luò)控制的混合小波神經(jīng)網(wǎng)絡(luò)盲均衡算法[J];電子學報;2011年04期
3 竇高奇;高俊;;適用于高階QAM系統(tǒng)的多模盲均衡新算法[J];電子與信息學報;2008年02期
4 郭業(yè)才;何龍慶;韓迎鴿;趙俊渭;;盲均衡技術(shù)在水聲信道均衡中的應(yīng)用進展[J];艦船科學技術(shù);2007年02期
5 姚祥光;周永權(quán);李詠梅;;二分微粒群協(xié)同進化優(yōu)化算法[J];計算機工程與應(yīng)用;2009年35期
6 于同泉;韋周芳;程彬彬;;ARC-FSK中程多載波水聲通信新技術(shù)[J];計算機仿真;2006年11期
7 李金明;趙俊渭;陸晶;;支持向量機初始化的常數(shù)模盲均衡算法仿真[J];計算機仿真;2008年01期
8 譚勝蘭;;模擬退火遺傳算法在網(wǎng)絡(luò)負載均衡中應(yīng)用研究[J];計算機仿真;2011年12期
9 王欣;杜康;秦斌;徐海軍;;基于果蠅優(yōu)化算法的LSSVR干燥速率建模[J];控制工程;2012年04期
10 許肖梅;;水聲通信與水聲網(wǎng)絡(luò)的發(fā)展與應(yīng)用[J];聲學技術(shù);2009年06期
,本文編號:1947946
本文鏈接:http://www.sikaile.net/kejilunwen/wltx/1947946.html