Study On Adaptive Equalization Algorithms Based On Wavelet Transform

In wireless digital communications , intersymbol Interference (ISI), which is responsible to a great part of bit errors, is rendered from time dispersion and phase offset. To restore the original information in the receiver, the adaptive equalization is capable of performing quite well in compensating the nonlinear distortion introduced by the channel. The purpose of an adaptive equalizer, placed in the path of the received signal, is to reduce the intersymbol Interference (ISI) as much as possible to maximize the probability of correct decisions. This thesis is to study the adaptive equalization algorithm based on wavelet transform, and Successfully applied it to nonlinear time-varying channels with severe distortion and additive Gauss noise (AGN), and eventually accelerate the equalizers'convergence speed and cut the bit-error rate, with less computation load.The content and innovation includes:①In order to resolve the problem of inter-symbol interference (ISI) because of limited bandwidth of the channel in wireless digital communications, the adaptive equalizer architectures and interrelated algorithms were deeply studied and improved by employing error feedback. The results of simulation showed that the error feedback equalizer is especially suitable in the multi-path fading environment.②Probed the wavelet equalization theory. Instead of thinking in time domain, I worked in the wavelet domain. Based on the multiresolution analysis, i discussed fast wavelet transform and Mallat's multiresolution decomposition. The Mallat algorithm , originally devised as a computationally efficient implementation, is more properly viewed as a multiresolution algorithm for the discrete wavelet transform.③Designed the wavelet equalizer. A new discrete orthogonal wavelet decision feedback equalizer based on error feedback was proposed. The feedforward filter is presented by a group of normalized orthogonal wavelets and their coefficients, while the feedback filter is taking error feedback and decision feedback. And the adaptive equalization algorithm is given. Computer simulations showed that error feedback was capable of converging to a lower value of mean square error by minishing auto-correlation of error signal. The accumulated bit error rate is almost still in trace phase. The wavelet equalizer showed its promptness, stability, simplicity and feasibility.④Revised the wavelet neural network equalizer and its adaptive algorithm. Propsed a new structure and algorithm of the wavelet neural network equalizer for nonlinear time-varying channels. The wavelet neural network is capable of forming the best approximation to any continuous nonlinear mapping up to an arbitrarly resolution.