A New Class of Codes with Error Localization: on improvement of decoding algorithms

This paper explores a novel reinterpretation of well-established cascade and generalized cascade coding structures, resulting in the development of a fundamentally new class of error localization codes (ELCs). The proposed constructon introduces a third constituent code responsible for localizing errors, which enables effective decoding with reduced complexity and improved performance. Three variations of the coding scheme are presented, including constructions with different outer code lengths and decoding strategies. Simulation results demonstrate that the proposed ELCs achieve near-maximum-likelihood performance while maintaining manageable decoding complexity. In particular, the use of short block codes and structured erasure patterns significantly improves the trade-off between reliability and computational efficiency. The paper concludes with a comparative analysis against polar and Reed–Muller codes and outlines promising directions for further research, including soft decoding strategies, joint coding-modulation schemes, and enhancements to inner code robustness.