Iterative decoding of serially concatenated convolutional codes

Serial concatenation of convolutional codes separated by an interleaver has recently been shown, through the use of upper bounds to the maximum likelihood performance, to be competitive with parallel concatenated coding schemes known in the literature as 'turbo codes'. The most important feature of turbo codes consists in their relatively simple, yet high performance, iterative decoding algorithm. The authors propose a new iterative decoding algorithm for serial concatenation, and show that the new coding scheme can yield a significant advantage with respect to turbo codes.