| Abstract: | Based on the property that correction coding is more sensitive to the number of errors than the number of erased code symbols, results of an analytical investigation to improve the performance of concatenated coding communications are presented. We propose a new tactic, ‘signal editing’, to estimate and erase unreliable data prior to the final outer decoding process. We show that signal editing has a capability to detect and delete individual erroneous digits that has not been achieved by other methods. As a result, system performance can be markedly improved. Specific recommendations for using the procedure are suggested and discussed. We illustrate how signal editing is implemented in the case of a convolutional code with Viterbi decoding. A new concatenated (or hybrid) coding technique is constructed integrating the code with interleaving and signal editing. Performance improvement is evaluated and analysed. In this study, signal editing is derived from the information available in the Viterbi decoding process. Not limited to convolutional code with Viterbi decoding, signal editing can be applied to other linear codes without any restriction. |
