An Improved Histogram Method for Calculating Extrinsic Information Transfer Functions

To remedy the defect of the current available methods used to calculate the Extrinsic Information Transfer (EXIT) Function; we propose an improved histogram method in this paper. It adopts an optimal histogram, which adjusts the bin width adaptively, to estimate the distribution of the LLR data. The optimal Probability Distribution Function (PDF) of the LLR data, in the sense of the integrated mean squared error (IMSE), is then estimated and from which the accurate EXIT function can be obtained. The proposed improved histogram method not only breaks through the application limitation of the algebraic average method, but also avoids the manual choosing of bin width in the original histogram method which does not guarantee the optimum. The simulation results show that the proposed method can get accurate EXIT function for different decoders, which provides a solid basis for the analysis and optimal design of the iterative processing system.