An adaptive threshold strategy for soft decision Viterbi decoding

Convolution codes are the most widely implemented type of forward error control (FEC) coding in wireless communication systems. The modest complexity and good performance have made the Viterbi algorithm the preferred decoding method for convolutional codes to overcome transmission errors. The availability of hardware and software for easy and efficient implementations contributed to the wide-spread applications of the various combinations of these two schemes.; Soft decision decoding of convolutional codes is known to have an improvement of about 2 dB when compared to hard decision decoding of similar codes in the Additive White Gaussian Noise (AWGN) channel. The corresponding performance enhance ments in Rician and Rayleigh fading channels are even more evident than in the AWGN channel. In the existing implementations of soft decision receivers quantizers have equi-spaced thresholds. In this thesis, to improve the bit error rate (BER) performance, two novel ideas are introduced in the framework of soft decision Viterbi decoding: (i) non-uniform spacing in quantizer thresholds; and (ii) adaptability in threshold levels according to signal-to-noise conditions in the channel.; Coherent BPSK and quaternary PAM transmission systems are taken as examples to illustrate the effectiveness of the adaptive multiple threshold quantizer. Some non equi-spaced threshold arrangement schemes are analyzed in order to devise an adaptive strategy that would optimize soft decision decoding in the flat-slow Rayleigh fading channels.; In mobile wireless communications, even a small BER improvement will have a considerable impact on the overall performance of the system. This small improvement in BER will save a significant amount of power over a period of time and extend the battery life. Therefore research in this field is of considerable interest to the academic community and industry.