| 摘要 |
<p>A constructing error-correcting codes based on 'expanded' encoding the product codes, is characterized that: after the input information is N-dimension product codes encoded, 'expanded' encoding operations are carried out on M-dimensional (1?=M?=N) directions therein. The 'expanded' encoding contains 'assembling' and 'shuffling' operations. Every group of 'expanded' encoding operations obtain a set of check symbols, new error-correcting codes are constructed by assembling the obtained M groups of check symbol and original N-dimension product codes. Also, the present invention discloses the corresponding iterative decoding method, wherein the iterative decoder is made up of a plurality of component decoders, every component decoder contain a soft-input-soft-output decoding unit and possible 'deassembling' and 'deshuffling' unit, which corresponds to the component code of the original product codes or 'expanded' encoding. At the same time, the interconnection between the respective component decoders can adopt a serial, parallel or combined form, wherein each stage of component decoder utilizes the outputs of all or part of component decoders which are added as a priori information to participant in this stage of decoding. Analysis and emulation all indicate that the error-correcting codes and decoding method of the present invention are capable of satisfy the design demand of a series of code rates, code lengths and performance requirements, simultaneously can employ varied embodiments according to actual situations, such as paralleling level, computation amount, decoding delay requirements, etc.</p> |
