Abstract
This dissertation proposes two channel-estimation based adaptive decision feedback equalizers (DFEs) to treat the estimation and equalization problems for multipath fading channels. The tap coefficients of such a fading channel are modeled as complex Gaussian stochastic processes. In the proposed design methods, the design procedure is divided into two steps. First, Kalman filtering algorithm which can achieve minimum variance estimator is developed to track the time-varying tap coefficients. Then based on the estimated tap coefficients and error covariance, two adaptive DFEs, which are constructed to minimize the mean square error (MSE) function in the frequency and time domain respectively, are designed to reconstruct the transmitted sequence.Since the information of channel dynamics (the state space equation of channel model) is employed to track the time-varying tap coefficients, the proposed adaptive DFEs exhibit a better reconstruction performance than the conventional adaptive equalizer. To demonstrate the superior performanceof the proposed adaptive DFEs, some illustrative simulations are presented. Furthermore, performance analysis about minimum mean square error (MMSE) achieved by the proposed design methods is also discussed in this study.