DVB-RCS: Efficiently quantized turbo decoder

Abstract

Turbo codes have been incorporated into many important wireless communication standards including the satellite return channel in DVB-RCS standards. According to their iterative nature, the computational complexity of turbo decoder is much higher than that of convolutional FEC decoders. From the hardware implementation point of view, the complexity can be reduced by using quantized decoder. For DVB-RCS turbo coding, there are many block sizes and different code rates. In order to realize the DVB-RCS turbo decoder efficiently, an algorithm should be developed for the best computation of quantization range for each code rate and at different signal-to-noise ratios. This paper investigates the decoder input quantization of low complexity decoding algorithm and proposes an algorithm for efficient decoder quantization by introducing a scaling factor into the decoding algorithm, aiming to achieve significant improvement in the hardware implementation of the decoder architecture.