Optimal channel equalization for filterbank transceivers in presence of white noise

Filterbank transceivers are widely employed in data communication networks to cope with inter-symbol-interference (ISI) through the use of redundancies. This dissertation studies the design of the optimal channel equalizer for both time-invariant and time-varying channels, and wide-sense stationary (WSS) and possible non-stationary white noise processes. Channel equalization is investigated via the filterbank transceivers approach. All perfect reconstruction (PR) or zero-forcing (ZF) receiver filterbanks are parameterized in an affine form, which eliminate completely the ISI. The optimal channel equalizer is designed through minimization of the mean-squared-error (MSE) between the detected signals and the transmitted signals. Our main results show that the optimal channel equalizer has the form of state estimators, and is a modified Kalman filter. The results in this dissertation are applicable to discrete wavelet multitone (DWMT) systems, multirate transmultiplexers, orthogonal frequency division multiplexing (OFDM), and direct-sequence/spread-spectrum (DS/SS) based code division multiple access (CDMA) networks. Design algorithms for the optimal channel equalizers are developed for different channel models, and white noise processes, and simulation examples are worked out to illustrate the proposed design algorithms