Design of hyper states for reduced-state sequence estimation

A new method for the design of hyper states for reduced-state sequence estimation of trellis-coded signals is proposed. As an example trellis-coded modulation (TCM) over channels with intersymbol interference (ISI) is used. The complexity of the super trellis resulting from a combination of the TCM and ISI trellises is reduced by an algorithm that applies set partitioning principles to the set of super states. This method allows to achieve asymptotic optimality at a low level of complexity. The new method is based on a well known algorithm for the computation of the minimum Euclidean distance. This algorithm Is extended to coded transmission with arbitrary trellis encoders and signal constellations. Simulations and comparisons with other methods show the superior performance of the proposed class of reduced-state sequence estimators.