Patent attributes
In an error correction method, a codeword is transmitted through a noisy communication channel and detected by a receiving device. An error detection code is then applied to the detected codeword to generate a syndrome. Where the syndrome is not all zero, the codeword is determined to contain some error. Accordingly, the method computes a set of potential error start positions for a plurality of error events based on a syndrome value corresponding to the syndrome. Next, a confidence value is computed for each of the plurality of error events at each of the potential error start positions in the refined set, and finally, a most likely error event in the detected codeword is corrected based on an error event and corresponding potential error start position having the highest confidence value.
