Orthogonal Wavelet Blind Equalization Algorithm Based On Artificial Immune System

Blind equalization techniques can improve system bandwidth utilization and avoid inter-symbol interference caused by multi-path transmission and bandwidth-constrained code, without increasing the bandwidth. Taking wavelet theory as a tool and making full use of the anti-self-correlation of wavelet transform, it can effectively speed up the convergence rate of blind equalization algorithm to transform signal with orthogonal wavelet.In recent years, the artificial immune system based on immune mechanism has received extensive attention of many scholars. It provides us so many evolutionary learning mechanisms, such as the noise patience, unsupervised learning, no negative examples, articulate and acquire knowledge, the accessibility of memory and forgetting of rarely used information. Aiming at disadvantages of constant modulus blind equalization algorithm, this paper deeply studies of the theory, algorithm and simulation of the artificial immune system based on blind equalization. The main work is as follows:1. Aiming at the shortcomings of constant modulus blind equalization algorithm, such as the slow convergence rate, large mean square error and easily to fall into local minimum points, this paper proposed orthogonal wavelet blind equalization algorithm based on artificial immune system. The algorithm takes the equalizer coefficient vector as antibodies, after a series of clone, variation and suppression, searches to the highest fitness value of the antibody, that is, the optimal coefficient of the equalizer, allows the weight vector out of local minima close to the global optimal; and uses of the orthogonal wavelet transform to improve the convergence of constant modulus blind equalization algorithm and reduce the mean square error. The simulation results show that the algorithm is effective, and optimizes orthogonal wavelet blind equalization algorithm.2. Introduce niche technology sharing mechanism into the artificial immune network algorithm to form a niche artificial immune network algorithm, and then use the algorithm to optimize the initial weight vector of the equalizer to improve the convergence speed, avoid premature convergence. The simulation results show that the algorithm has a faster convergence rate and smaller steady state error and better anti-interference performance.3. Introducing shape-space artificial immune regulatory network algorithm into the orthogonal wavelet constant modulus algorithm, which is a combination of morphological space model and the idiotypic network regulation theory, wavelet transform constant modulus algorithm based on shaped-space artificial immune regulation network is proposed to improve the algorithm convergence speed and reduce the mean square error. The simulation results show that the algorithm has good optimization results.