Adaptive decoding method for double‐encoding systems

Abstract

AbstractA general decoding method for doubleencoding systems has been proposed by Imai and Nagasaka, wherein decoding state information obtained in the first stage is utilized fully. This method has a better decoding performance than conventional ones and implementation is easy. However, because this method is based on the probabilistic characteristic of communication channel, its decoding performance can deteriorate when BER of communication channel is varied. This paper proposes an adaptive decoding method which follows the variation of probabilistic characteristic of communication channel. In this method, channel state information can be obtained in the first stage, then the second decoding stage is performed appropriately to that channel state information. With this proposed method, high quality communication is achieved when the communication channel is in good condition. Moreover, even if some quality degradation is unavoidable, communication that is as good as possible is achieved even when the communication channel is in a bad condition. Since an enormous amount of computation is necessary to determine the optimum adaptive decoding system, an adaptive decoding method using some approximation is studied. This method is suitable for a channel where transitions of the channel states can be modeled as a Markov process. Numerical comparisons show that the proposed method achieves a better decoding performance.