Immune Optimization Algorithm And The Application In Orthogonal Wavelet Blind Equalization

With the need to high rate underwater acoustic communication is more and more urgent, one of the main obstacles is the iner-symbol interference caused by the effect of multipath propagation, the goal to get the data and the reliability has become a research focus in the underwater acoustic communication. Blind equalization does not need training sequence to track channel changes to be possible to effectively save bandwidth. The performance of the Multimodal function optimization of clonal selection algorithm and the decorrelation ability of wavelet transform. Overcome to disadvantages of low convergence rate and local minima in traditional constant modulus blind equalization algorithm(CMA). The article has conducted the in-depth study to immune clonal selection algorithm. The main contributions are as follow:1Analyzing orthogonal wavelet transform blind equalization algorithm based on the optimization of Immune clonal algorithmIn the traditional constant modulus algorithm of blind equalization, the method of searching needs to derivable as the main conditions is a gradient descent search method. But the method is easy to fall into local convergence. Immune clonal selection algorithm is introduced to the blind equalization algorithm, it is full use of the multimodal function optimization of the clonal selection algorithm to find the global optimal solution of the equalizer. Aiming at the decorrelation of the orthogonal wavelet transform can speed up the convergence rate of the equalizer, and can quickly find the global optimal solution of the equalizer weight vector. Simulation results show that it can improve the performance of the proposed algorithm effectively.2Analyzing orthogonal wavelet multi-modulus blind equalization algorithm based on immune optimization of support vector machineSince SVM blind equalization search for the optimize weight vector through the linear programming, the search method show a good performance. But in the SVM parameters settings for final classification accuracy have a major influence in the process of constructing. In order to make the SVM has higher accuracy and better generalization ability, it's need to select the reasonable parameter. Using characteristics of the global search of immune clonal selection algorithm to optimize the parameter. In order to overcome the slow convergence rate, big steady-state and local convergence error of CMA when balanced the high order QAM. An orthogonal wavelet transform multi-modulus blind equalization algorithm of support vector machine is proposed, and use the optimized by immune algorithm of SVM to initialize equalizer, it's effectively to avoid the local convergence of the algorithm.3Analyzing orthogonal wavelet multi-modulus blind equalization algorithm based on the optimization of improved Immune clonal algorithmIn the standard immune clonal algorithm, the antibodies scale, clonal selection scale and high frequency variation rate values produce very big effect of the algorithm. Therefore, the algorithm based on the antibodies scale, clonal selection scale and high frequency mutation probability of the affinity self-adaptive in evolution process of immune system, and the combination of the stochastic search with the uncertain evolution search. And applied to blind equalization algorithm to overcome the defects of standard immune clonal algorithm.By using SEI to whitening the input data of equalizer, the simulation results show that the proposed algorithm has better balanced performance. The standard immune algorithm of cloning have stronger dependence on initial solution, late in the iteration it's easy to appear stagnant phenomenon and fall into the equilibrium state of local optimal, it leads to low precision of the search and produced the local optimal solution. Introduced MPCSA to orthogonal wavelet multimode blind equalization algorithm, It effecttively improves the performance of the proposed algorithm by computer simulations and theoretical analysis. Figure [32] Table [4] Reference [104]...