Analysis of the Sufficient Path Elimination Window for the Maximum-Likelihood Sequential-Search Decoding Algorithm for Binary Convolutional Codes

TL;DR

This paper analyzes the optimal early-elimination window in the maximum-likelihood sequential-search decoding algorithm for binary convolutional codes, showing that a small window suffices for near-ML performance, reducing complexity.

Contribution

It introduces an early-elimination strategy for MLSDA and derives theoretical bounds on the window size needed for negligible performance loss, supported by simulations.

Findings

Early-elimination window is approximately twice the constraint length for rate 1/2 codes.

For rate 1/3 codes, the window reduces to the constraint length.

The proposed method significantly reduces computational complexity and memory usage.

Abstract

A common problem on sequential-type decoding is that at the signal-to-noise ratio (SNR) below the one corresponding to the cutoff rate, the average decoding complexity per information bit and the required stack size grow rapidly with the information length. In order to alleviate the problem in the maximum-likelihood sequential decoding algorithm (MLSDA), we propose to directly eliminate the top path whose end node is $\Delta$-trellis-level prior to the farthest one among all nodes that have been expanded thus far by the sequential search. Following random coding argument, we analyze the early-elimination window $\Delta$ that results in negligible performance degradation for the MLSDA. Our analytical results indicate that the required early elimination window for negligible performance degradation is just twice of the constraint length for rate one-half convolutional codes. For rate one-third convolutional codes, the required early-elimination window even reduces to the constraint length. The suggestive theoretical level thresholds almost coincide with the simulation results. As a consequence of the small early-elimination window required for near maximum-likelihood performance, the MLSDA with early-elimination modification rules out considerable computational burdens, as well as memory requirement, by directly eliminating a big number of the top paths, which makes the MLSDA with early elimination very suitable for applications that dictate a low-complexity software implementation with near maximum-likelihood performance.