Study On The Adaptive Equalization Technology In Nonlinear Channel

In the fast mobile circumstance with high data rate, the communication channel nonlinearity will become the main factor affecting the bit error rate, when only the nonlinear equalization technology can remove the ISI effectively. So, the research on the nonlinear channel equalization is now a hot spot in the high-rated wireless communication area.This paper is performed mainly on the adaptive equalization technology in the nonlinear channel, based on summarizing the present study on nonlinear adaptive signal processing and nonlinear channel equalization domestically and abroad. By using adaptive filtering and neural network technology, novel ways on how nonlinear adaptive filter (or neural network) can be applied into the nonlinear equalization are discussed and several novel nonlinear equalizers are proposed, which have shown better equalization performance than the general ones.The main results of this paper are as follows:1. A nonlinear Volterra filter with a product coupling item is applied in the channel equalization and its modified structure is given, which has a very simple structure but with good nonlinear equalization performance.2. An over-sampled blind equalization strategy based on the noise suppression in DCT domain is studied. Two results are reached that the blind FSE can handle the nonlinear equalization problem to some degree and that through the DCT filtering the equalizer's anti-noise performance can be improved effectively.3. The decision-feedback structure of the neural adaptive filter is studied and a novel bilinear nonlinear neural DFE is proposed, which has less mean square error. And, the complex algorithm of the neural DFE is given, which can effectively compensate the phase and amplitude loss in the complex channel.4. A novel neural Chebyshev orthogonal polynomial adaptiveequalizer is proposed, which is stable and has stronger nonlinearity-processing ability than the general nonlinear equalizers.