A simplification of the Log-MAP algorithm for turbo decoding

Abstract

The Log-MAP algorithm and its modifications were proposed to simplify the VLSI implementation of turbo decoding. This work presents a new way to simplify the Log-MAP algorithm, called Shift-Log-MAP as it takes place of log and exp operation with shift. The correction term in the Log-MAP algorithm was transformed to linear function by use of Taylor equation at the slope of 1/2, 1/4, etc. It allows for efficient ASIC implementation and almost has the same performance as the Log-MAP algorithm.