An adaptive hybrid ARQ scheme based on exponentially weighted moving averages for slowly time-varying communication channels

Abstract

An adaptive hybrid ARQ scheme for slowly time-varying channels that aims to maximize the average throughput under a given reliability constraint is presented. Punctured convolutional codes and the Viterbi algorithm with the Yamamoto-Itoh retransmission protocol are employed for encoding and decoding, respectively. A procedure based on exponentially weighted moving averages is developed to detect channel state changes.<>