Comparison and analysis of ECSS based turbo decoder stopping rules

Turbo decoding is an iterative process in which the extrinsic information is computed based on Log-Likelihood Ratio (LLR) and the reliability of data is improved as we increase the number of iterations. The total number of decoder iterations required to obtain reliable data depends upon the physical channel characteristics; the noise level and fading magnitude. Hence, it is important to develop some criteria to stop the decoding process once there are no further improvements in the extrinsic information in order to save the computational power and increase the decoding speed. In this paper, we have compared and analyzed the performance of several different existing stopping rules (Early Termination) for turbo decoder used for the Satellite Telemetry link. We have also proposed a modified early termination rule (MCDR) based on the combination of sum-reliability and min-reliability of extrinsic information which yields improved system performance. We have performed the simulations and plotted the average number of iterations and bit-error-rate (BER) curve for these stopping rules and discussed the buffering requirement associated with them.