Maximum likelihood sequence estimation of CPM signals transmitted over Rayleigh flat-fading channels

Equalization of Time-Varying Channels

“…As mentioned earlier, Lodge and Moher [LM90] were one of the first papers to propose a near-optimal trellis-based implementation of noncoherent MLSD. In particular, they proposed to use the Viterbi algorithm with per-branch linear prediction for MLSD of CM signals with ARMA time-selective channels.…”

“…Dai and Shwedyk [DS94] proposed similar near-optimal trellis-based implementations of noncoherent MLSD for general signal alphabets and ARMA doubly selective channels, using per-branch Kalman filtering. Yu and Pasupathy [YP95] then extended [LM90] to general signal alphabets and ARMA doubly selective channels. The latter technique was extended further to carrier-frequency-offset Rician channels by Hart and Taylor in [HT98].…”

“…This idea seems to have been first proposed in Morley and Snyder [MS79] using continuous-time filtering. Lodge and Moher [LM90] considered discretetime filters and realized that, if the observations are first whitened (which requires only a bank of LTI filters), then the metric calculation simplifies in a way that eliminates the need for Kalman filtering. This latter approach is known as the "innovations" approach.…”

Joint Estimation and Decoding of Space-Time Trellis Codes

“…Then the proposal distribution for the delayed sampling method can be derived similarly as the one of the non-delay case, and it results in p s t |s The term inside the summation in (22) can be readily evaluated. The weights of the samples are obtained similarly as in (20). Based on these derivations, an algorithm was developed for the online estimation of user data using the delayed sample method when the state vector can be partitioned into two parts.…”

“…In communications, SIS methods have been used for blind detection of user data [16,17,18]. Because there are no mathematical approximations of the model, SIS methods are found to perform better than traditional methods like the Extended Kalman Filter (EKF) [19,20] or the Gaussian Sum Filter (GSF) [21]. It should be noted, however, that the sampling based methods use approximations too in that they approximate relevant densities with samples from these densities.…”