A code-aided adaptive equalizer using soft decision-directed algorithm and convolutional coding for fading channels

Abstract

This thesis investigates a modified soft output Viterbi algorithm applied to a code aided adaptive equalization system for mobile communications. Adaptive equalization is a necessary signal processing technique in a frequency selective fading environment. It is usually employed in a receiver. This way, the receiver periodically tracks the time varying channel and minimizes the intersymbol interference (ISI) on the received signals accordingly. In general, it is impossible to totally remove the time-varying ISI from the received signal. Hence errors still occur, though in a reduced manner, after the equalization process. Coding techniques can be employed to tackle the error residues. A conventional receiver for mobile communication systems makes use of adaptive equalization and coding techniques to help extract the intended transmitted signal from a multipath fading channel. In high speed mobile communication systems, such as in transmitting streaming video and multimedia services, a good bit error rate (BER) performance is required in order to provide an acceptable quality of service. To maintain a good transmission performance, it is crucial to improve the performance of the mobile receiver so that the resulting bit error rate can satisfy the BER requirement. We focus on the improvement in the equalization structure in order to enhance the equalizer adaptation process with the aide of coding techniques. We investigate ways to enhance the adaptive process so that the number of errors in the equalized signal can be minimized. Based on a two-stage, code-aided, adaptive equalization scheme, which has been shown to be effective in mitigating the distortions in the transmitted signals from the intersymbol interference channel, we propose to impose a modified version of the soft output Viterbi algorithm in the second stage equalizer to provide more precise channel information to the equalizer. This leads to a receiver with a better bit error rate performance. We also investigate non-linear decision functions applied in equalization to enhance the tap coefficients updating process of an equalizer, and focus on a non-linear function we propose. The proposed non-linear decision function enables an equalizer to give smaller responses to weaker signals, which are presumably unreliable, and to give larger responses to comparatively stronger signals in a non-linear manner. This can suppress those weak signals which are not conveying reliable information to the equalizer and which actually adversely affect the equalizer adaptation process. We evaluate the performance of the new schemes in Rayleigh fading environments. Computer simulations are carried out to verify the performance of the proposed equalizer schemes in Rayleigh fading channels for different fade rates. In comparing them with the conventional equalization, the two proposed approaches are shown to provide promising improvements.