Sequence estimation scheme for narrowband digital FM signals with limiter-discriminator detection under fading environment

A communication scheme using binary FM with noncoherent limiter-discriminator detection has been well known. Up to now, the bit error rate improvement at the receiver side has been carried out through the bandwidth optimization of the IF filter, the decision feedback equalization (DFE) or simple maximum likelihood sequence estimator (MLSE). This channel is inherently the intersymbol interference (ISI) channel due to the pre-modulation baseband filtering as well as the narrow band IF filtering. So the sequence estimation scheme using the Viterbi algorithm can be applied successfully, although the channel is not additive white Gaussian and maximum likelihood in the strict sense. In this paper, through computer simulations, we examine the actual BER improvement of the sequence estimation scheme using the Viterbi algorithm especially for MSK and GMSK signals. We consider static AWGN, frequency non-selective Rician and Rayleigh fading channels. Consequently, very large estimation gains are obtained compared to the conventional DFE or MLSE detector for AWGN and Rician fading channels. This scheme can be implemented only by adding the signal processing part to the final stage of the receiver and seems to be very useful for any applications including satellite mobile channels