Estimating the Bit Error Rate in Digital Communication Systems

W. Song and W. Chiu (Taiwan)


bit error rate, importance sampling, Monte Carlo simulation, mean squared error


The bit error rate (BER) has been considered as an important performance measure for digital communication systems. Simulation approaches are recommenteded to estimate the BER when the threshold in the digital communication systems is random or the noise process is not a white noise process. Importance sampling (IS) is one variance reduction technique often used to increase the efficiency of simulation experiment. Shih and Song (1995) showed that a mixed biasing distribution, which combines a truncated Gaussian and the Uniform distribution, performs better that the commonly used truncated Gaussian distribution when the signal or the threshold setting follows the uniform distribution. This paper continues to investigate the performance of using a mixed biasing distribution in importance sampling for estimating the bit error rate (BER) via Monte Carlo simulation.

Important Links:

Go Back