New results on the search for convolutional self-doubly orthogonal codes suitable for iterative threshold decoding

We present new results on the search and determination of wide-sense convolutional self-doubly orthogonal codes (CSO²C-WS) which can be decoded using a simple iterative threshold decoding algorithm without interleavers. For their iterative decoding, in order to ensure the independence of observables over the first two iterations without the presence of interleavers, these CSO²C codes must satisfy some required orthogonal properties. The error performance of CSO²C codes, depends essentially on the number of taps, J, of the code generators, but not on the code memory length. We present a code searching technique, based on computer search heuristic algorithms, which has yielded the best known CSO²C-WS codes. Error performance simulation results are also provided. The construction method for CSO²C-WS has provided new codes with the shortest memory length having J ≤ 30.