Importance sampling technique for multi-user CDMA-like systems in discrete-time multipath fading environments

Monte-Carlo simulations of multi-user systems using code division multiple access (CDMA) are by nature rather inefficient because of the high sampling rates implied in the processing of the spread signal. One promising way to improve simulation efficiency is to make use of improved importance sampling (IIS). This technique is based on a translation of the original probability density of the noise process in order to generate more errors and weighting these errors afterwards. This yields a BER estimate with the same variance for a reduced number of simulation runs. In this paper, the application of IIS is made to CDMA-like multi-user systems over multipath fading channels. In contrast to classical CDMA, these broadband systems achieve a certain degree of orthogonality between the users thus allowing an enhanced spectral efficiency