On the design of rate-compatible serially concatenated convolutional codes

A powerful class of rate-compatible serially concatenated convolutional codes (SCCCs) has been proposed based on minimizing analytical upper bounds on the error probability in the error floor region. In this paper, this class of codes is further investigated by combining analytical upper bounds with extrinsic information transfer chart analysis to improve performance in the waterfall region. Following this approach, we construct a family of rate-compatible SCCCs with low complexity and good performance in both the error floor and the waterfall regions over a broad range of code rates. The performance of the proposed codes is found to compare favorably to standard SCCCs and PCCCs. The proposed codes outperform standard SCCCs and have a similar performance as more complex PCCCs. The proposed codes perform particularly well for high code rates.