Signal separation using fractional sampling in multiuser communications

Described is a decorrelating filter that separates signals in a multiuser environment by relying on the relative delays to be sufficiently distinct. The input signal is fractionally sampled to allow for the differentiation of the user delays. Both zero-forcing and minimum mean-square-error versions of this filter are derived and probability of error expressions are given. The algorithm successfully separates unknown digital signals by using the known received pulse shapes and the symbol rate. A delay-division multiple-access (DDMA) scheme based on this signal decorrelator is proposed that will allow signals to be transmitted without spreading the signal spectrum. It is shown that in a noisy environment the signal-to-noise ratio penalty for DDMA can be severe yet the bandwidth advantage is very pronounced compared to spread-spectrum systems and is similar to other bandwidth efficient schemes. The performance of a code-division multiple-access (CDMA) system using this signal decorrelator is also given. The decorrelator can be used as a blind multiuser detector or as a preprocessor to enhance the performance of a CDMA system.