On low complexity stack decoding of convolutional codes

Abstract

This letter presents techniques for improving the distribution of the number of stack entries, for stack sequential decoding over hard quantized channel, with emphasis on high rate codes. It is shown that, for a class of high rate b/(b+1) codes, a table-based true high rate approach can be easily implemented for obtaining decoding advantage over the punctured approach. Modified algorithms, which significantly improve the distribution of the number of stack entries and decoding time, are proposed for rate 1/N codes and high rate b/(b+1) codes.