Statistical evaluation of the error rate of the Fano and stack algorithm decoders

Abstract

The estimation of the bit error probability is fundamental to the performance evaluation of today's complex digital communication systems. Unfortunately, from a computational point of view the conventional Monte Carlo (MC) approach proves to be extremely difficult, if not impossible, for simulating such systems. As a result, it is highly desirable to devise MC methods which retain the ability to simulate complex systems, yet require substantially fewer simulation runs to estimate small probabilities. A new quick MC technique for estimating the bit error rate (BER) of sequential decoders is presented. It is shown that the new technique is extremely efficient from a computational point of view and can be used to accurately estimate the BER for specific time invariant convolutional codes.<>