Abstract
Subtrellises for low-weight codewords of binary linear block codes have been recently used in a number of trellis-based decoding algorithms to achieve near-optimum or suboptimum error performance with a significant reduction in decoding complexity. An algorithm for purging a full code trellis to obtain a low-weight subtrellis has been proposed by H.T. Moorthy et. al. This algorithm is effective for codes of short to medium lengths, however for long codes, it becomes very time consuming. This paper investigates the structure and complexity of low-weight subtrellises for binary linear block codes. A construction method for these subtrellises is presented. The state and branch complexities of low-weight subtrellises for Reed-Muller codes and some extended BCH codes are given. In addition, a recursive algorithm for searching the most likely codeword in low-weight subtrellisbased decoding algorithm is proposed. This recursive algorithm is more efficient than the conventional Viterbi algorithm.
| Original language | English |
|---|---|
| Pages (from-to) | 2095-2102 |
| Number of pages | 8 |
| Journal | IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences |
| Volume | E80-A |
| Issue number | 11 |
| Publication status | Published - Jan 1 1997 |
| Externally published | Yes |
Keywords
- Binary linear code
- Low weight
- Recursive decoding algorithm
- Subtrellis diagram
- Subtrellis search
ASJC Scopus subject areas
- Signal Processing
- Computer Graphics and Computer-Aided Design
- Electrical and Electronic Engineering
- Applied Mathematics