Punctured Parallel Viterbi Decoding Of Long Constraint Length Convolutional Codes

Abstract

This paper considers puncturing applied to the parallel Viterbi decoding of long constraint length convolutional  codes involving partitioning of the basic code. It is shown that the combination of the two techniques of parallel  Viterbi decoding and puncturing provides a considerable reduction in decoding complexity, which reflects itself as a gain in decoding speed, but without sacrificing the code rate. The error performance of the basic constraint length 8 rate ½ convolutional code is compared with those of (1) the rate ¼ constraint length 8 parallel code partitioned into two rate ½ constraint length 4 codes and (2) the punctured parallel code which restores the code rate back to ½. It is found that, for the selected puncturing map, a degradation of about 0.8 dB in the error performance of the punctured partitioned convolutional code is obtained as compared to the performance of the nonpunctured partitioned code with soft decision decoding.

Keywords: Convolutional codes, Parallel Viterbi, Partitioning, Puncturing.