Published December 6, 2019 | Version v1
Journal article Open

Revisiting the Max-Log-Map algorithm with SOVA updates rules: new simplifications for high-radix SISO decoders


This paper proposes a new soft-input soft-output decoding algorithm particularly suited for low-complexity highradix turbo decoding, called local soft-output Viterbi algorithm (local SOVA). The local SOVA uses the forward and backward state metric recursions just as the conventional Max-Log MAP (MLM) algorithm does, and produces soft outputs using the SOVA update rules. The proposed local SOVA exhibits a lower computational complexity than the MLM algorithm when employed for high-radix decoding in order to increase throughput, while having the same error correction performance even when used in a turbo decoding process. Furthermore, with some simplifications, it offers various trade-offs between error correction performance and computational complexity. Actually, employing the local SOVA algorithm for radix-8 decoding of the LTE turbo code reduces the complexity by 33% without any performance degradation and by 36% with a slight penalty of only 0.05 dB. Moreover, the local SOVA algorithm opens the door for the practical implementation of turbo decoders for radix-16 and higher.



Files (922.7 kB)

Additional details

Related works

Is supplemented by
hal-02332503 (hal)


EPIC – Enabling Practical Wireless Tb/s Communications with Next Generation Channel Coding 760150
European Commission