Blind Equalization Algorithms Based On Modified Genetic Algorithm And Wavelet Transform

Inter-symbol interference caused by distortion and aberration of underwater acoustic is a major factor affecting the quality of communication. In order to improve the reliability of rate communication and transmission rate, it’s necessary to equalize the signal at the receiving end. Needing to repeat sending the training sequence, traditional adaptive equalization techniques will occupy the bandwidth of underwater acoustic which is very narrow, and reduce the communication efficiency. Blind equalization technology which does not need to transmit the training sequence can conserve bandwidth effectively as well as improve communication efficiency. With the development of modern underwater acoustic high-speed communication, it has great significance to analyze and research the blind equalization algorithms.Aiming at the disadvantages such as slow convergence rate, large steady-state error and the local convergence of Constant Modulus Algorithm (CMA), the Modified Genetic Algorithm (MGA) and Wavelet Transform (WT) and other means is used to optimize the blind equalization algorithm in this paper. The main contributions are as follows:1. Proposing orthogonal Wavelet Transform Constant Modulus Algorithm based on Genetic Algorithm (GA-WTCMA). To correct defects of orthogonal wavelet constant modulus algorithm, the GA-WTCMA is proposed. In the algorithm, the weight vector coefficient of blind equalizer is regarded as individuals of GA’s population. The fitness function is proposed according to the cost function of WTCMA. Because of the global search feature of blind equalization algorithm, GA-WTCMA can avoid the local convergence efficiently.2. Proposing orthogonal Wavelet Transform Constant Modulus Algorithm based on Modified Hybrid Genetic Algorithm (MHGA-WTCMA). Aiming at the disadvantages such as low operation efficiency and premature convergence and so on, the Modified Genetic Algorithm (MGA) is proposed. To constitute MHGA-WTCMA, WTCMA is embedded into the genetic algorithm which operates alternately with another one.3. Proposing the orthogonal Wavelet Transform and Spatial Diversity combined Blind Equalization Algorithm based on Parallel Genetic Algorithm (PGA-SD-WTCMA). The weight vector coefficient of blind equalizer in each branch of SD-WTCMA is regarded as sub-population to be optimized by parallel genetic algorithm. The sub-populations will also be migrated with each other regularly. The proposed algorithm can speed up the convergence rate and improve the accuracy of the solution.4. Proposing the orthogonal Wavelet Transform and Spatial Diversity combined Blind Equalization Algorithm based on Adaptive Parallel Genetic Algorithm (APGA-SD-WTCMA). Aiming at using fixed parameters in each sub-population of parallel genetic algorithm, as well as different channel characteristics of each branch in SD-WTCMA, the Adaptive Parallel Genetic Algorithm (APGA) which can set parameters adaptively according to the different environment and evolution period is proposed. With adaptive crossover and mutation operators and adaptive migration operator, APGA can set parameters more reasonability. Optimized by APGA, APGA-SD-WTCMA has faster convergence speed and better global convergence.