APP Convolutional Decoding with Transition-Based Systematic Channel Estimation

Abstract

This paper presents a novel formulation for a posteriori probability (APP) decoding of systematic convolutional codes. The convolutional encoder and decoder are constructed to enable transition-based channel estimates to be embedded into the APP calculations. The result is joint channel estimation and decoding. The new decoder is targeted to systematic codes in flat-fading environments although the formulation may be extended for frequency-selective channels or even non-systematic codes with the penalty of additional complexity. In contrast to per-survivor processing for Viterbi decoding, the approach here does not rely on tentative decisions from survivor paths, channel estimation filter coefficients can be pre-calculated, and the APP decoder delivers soft decisions.