Low-complexity Chase decoding of algebraic-geometric codes using Koetter's interpolation
Siyuan Wu, Li Chen, Martin Johnston · 2016
Algebraic-geometric (AG) codes have long been considered as a possible candidate to replace Reed-Solomon (RS) codes. However, their decoding remains complex and infeasible to implement. Addressing this challenge, our paper proposes a low-complexity Chase (LCC) decoding algorithm for the most popular class of AG codes - Hermitian codes. The LCC decoding is realised by formulating decoding test-vectors, which allows Koetter's interpolation to be performed for common and uncommon elements. This reduces redundant computations and also removes the need to calculate the corresponding coefficients of a Hermitian curve, thus facilitating message recovery. Our simulation results show that significant coding gains can be achieved over the conventional Koetter-Vardy (KV) soft decoding algorithm, but with a much lower computational cost. Moreover, we also show that in comparison with RS codes of a similar length, Chase decoding has a more significant impact on enhancing the performance of Hermitian codes.