Fast estimation of the Frame Error Rate for error-correcting codes by the generalized error impulse method

In this paper, we describe a new method to determine the Frame Error Rate (FER) for error-correcting codes in the presence of additive white Gaussian noise (AGWN) under a maximum-likelihood (ML) hypothesis. This algorithm is a generalization of the impulse method of Berrou et al. We compare our method with the standard Monte Carlo method for the estimate of low frame error rates for the extended Golay code. We observe large speedups concerning simulation time. We also experiment this method in the case of turbo product codes. The estimated performance obtained by the new method turns out to be very close to the exact performance obtained by the Monte Carlo method.