Soft-output sequential algorithm for signal estimation over frequency selective fading channel

The performance of the receiver in digital mobile radio systems can be improved if the equalizer delivers the reliability information to the channel decoder. The soft-output Viterbi algorithm (SOVA) is well known for this application. However, for long channel memory, SOVA becomes impractical due to the exponential increase in computational complexity. On the other hand, the sequential estimation algorithms can be applied since the computational complexity is independent of the channel memory. This paper presents a soft-output sequential algorithm (SOSA) for estimation of signal transmitted over frequency selective fading channel. SOSA generates the soft-output information by applying the likelihood post-processing technique after each decoded block of symbols. Simulation results show that SOSA can achieve performance gain over the conventional maximum-likelihood sequence estimation (MLSE) and that its performance approaches SOVA as the signal-to-noise ratio increases. Other possible applications of SOSA are also mentioned.