Nonlinear communication system simulation via conditional importance sampling

The basic concepts and advantages of the conditional importance sampling (CIS) technique are presented. By conditioning on random sources, this technique simplifies the problem of biasing to lower dimensionalities and, because of its adaptive nature, its biased PDF can be brought closer to the optimum solution. The CIS technique is applied to the simulation of bit error rate in nonlinear digital communication systems. The results obtained from a performance comparison of the CIS with an improved importance sampling technique using a system with two additive white Gaussian noise sources are presented