Abstract: A method for estimating the BER for telecommunication systems, particular those characterized by signals having high crest factors or causing large inband nonlinear distortions. The set of signals used by the system is divided into subsets according to a characteristic such as signal crest factor, and a BER estimation method is chosen for each subset. Signals causing a large BER are simulated more efficiently using a Monte Carlo simulation, while low BER estimations more efficiently use a quasi-analytical method. The method results in improved accuracy because the noise contribution in quasi-analytical methods can be better approximated for signals having a small crest factor range, and drastically reduces the number of experiments, measurements or simulations which are needed to obtain an accurate BER estimation, as compared to standard Monte Carlo techniques.

Abstract: A method for estimating the BER for telecommunication systems, particular those characterized by signals having high crest factors or causing large inband nonlinear distortions. The set of signals used by the system is divided into subsets according to a characteristic such as signal crest factor, and a BER estimation method is chosen for each subset. Signals causing a large BER are simulated more efficiently using a Monte Carlo simulation, while low BER estimations more efficiently use a quasi-analytical method. The method results in improved accuracy because the noise contribution in quasi-analytical methods can be better approximated for signals having a small crest factor range, and drastically reduces the number of experiments, measurements or simulations which are needed to obtain an accurate BER estimation, as compared to standard Monte Carlo techniques.