Browse > Article

Performance Evaluation of the M-algorithm for Decoding Convolutional Codes  

천진영 (서강대학교 전자공학과 디지털전송연구실)
최규호 (서강대학교 전자공학과 디지털전송연구실)
성원진 (서강대학교 전자공학과 디지털전송연구실)
Abstract
The M-algorithm for decoding convolutional codes can significantly reduce the complexity of the Viterbi algorithm by tracking a fixed number of survivor paths in each level of the decoding trellis. It is an easily-implementable algorithm suited for real-time processing of high-speed data. The algorithm, however, generates a sequence of catastrophic errors when the correct path is not included in the set of survivor paths. In this paper, the performance of the M-algorithm obtained from using various decoding complexity levels, frame lengths, and code constraint lengths is presented. The performance gain is quantified when the algorithm is used in conjunction with codes of increased constraint length. In particular, it is demonstrated the gain from the increased code free distance overcompensates the loss from the correct path being excluded from the survivors, when the frame length is short to moderate. Using 64 survivor paths, the signal-to-noise ratio gain obtained by increasing the constraint length from K=7 to K=9, 11, 15 is respectively 0.6, 0.75, and 08dB, when the frame of length L=100 has the frame error rate of 0.01%.
Keywords
Citations & Related Records
Times Cited By KSCI : 1  (Citation Analysis)
연도 인용수 순위