Applicability of importance sampling for Rayleigh fading mobile radio channels

Estimating the bit error rate (BER) of complex digital communication systems has often been realised by using the Monte Carlo method. For very low BER, this method requires a high computational effort. Therefore, a lot of developments in theory and application of importance sampling (IS) techniques have been done especially for AWGN channels or time-varying uniformly distributed channels. In this paper we discuss the applicability of several IS techniques for the simulation of Rayleigh fading mobile radio channels. By using the system threshold characteristic (STC) we first make an analytical statement of the IS efficiency for an AWGN and a frequency-nonselective Rayleigh fading channel. It is shown that due to the Rayleigh distribution the efficiency is reduced. Furthermore, the so called overbiasing effect is graphically explained and simulation results illustrate that the theoretically possible efficiencies cannot be reached in practice because of this statistic effect. Finally, the influence of time-selectivity is analysed