Abstract: In an asynchronous DS-CDMA system of multiple users, there are inter-symbol interference(ISI)，multiple access interference(MAI) and external interference(EI). According to function space thcory,decomposing the narrowband external interference on a function space with finite dimensions, regarding each basis of the function space as a vector of the mixing matrix, and taking the coefficient of corresponding basis as one of sources, we proposed a BSS-based method, requiring the received signal to be over-sampled, which can realize both blind equalization and blind multi user detection in the case that all these three interferences coexist simultaneously. The simulation result indicates that this method performs better than RAKE with perfect equalization, especially in lower SIRs.

Key words: DS-CDMA, Blind multiple user detection, Blind equalization, External interference cancelation, Blind source separation, Maximum non-gaussianity