Multiuser Detection for Cdma Systems with Nonlinear Modulation

Most past research on multiuser detection has focused on linearly modulated systems. This paper considers the more general problem of multiuser detection with nonlinear modulation. Our goals are to (a) gain fundamental insight into the capabilities of multiuser detection in such a setting, and (b) to provide practical algorithms that perform better than conventional methods. For the rst goal, we provide performance benchmarks by considering coherent detection using the maximumlikelihood receiver (ML), which requires knowledge of all transmission parameters, and noncoherent detection based on the generalized likelihood ratio test (GLRT), which does not require knowledge of user amplitudes and phases. The asymptotic performance of both detectors, as noise level vanishes, is characterized. Since the complexity of the ML and GLRT detectors is exponential in the number of users, for our second goal, we propose the sequential decision projection (SDP) detector, which has quadratic complexity in the number of users. It is shown that the SDP detector has nonzero asymptotic eeciency, provided that the users' powers are suitably disparate. While the work here applies to coded multiuser systems in general, a speciic application of current interest is the mobile-to-base \uplink" for the IS-95 code division multiple access (CDMA) based standard for digital cellular telephony, which employs orthogonal modulation.