Multi-user Detection for Signals Having Different Symbol Rates

This paper proposes a multi-user detector that detects signals having different symbol rates, within a reasonable level of computational complexity. The proposed multi-user detector successively cancels undesired signals for each user in a received signal. In the proposed method, the undesired signals are band-limited using a band-pass filter for a desired signal, and the desired signal is detected from the band-limited signal. A replica of the desired signal is generated based on soft-output from a channel decoder and subtracted from the received signal in order to remove its influence on the signal detection of other users. Then, another user is detected by using the residual signal obtained by subtracting the replica signal. A performance comparison of the proposed method to a multi-user detector based on the maximum likelihood sequence estimation (MLSE) is presented based on the computer simulations. The computer simulation results show that the proposed method can separate the signals effectively, and that its BER performance for a narrow-band signal approaches that of the MLSE-based multi-user detector when the symbol rate ratio of the wide-band signal to the narrow-band signal becomes large