Low-complexity Joint Data Detection and Channel Estimation in Time-varying Flat-fading Channels within the Sage Framework

We derive a multi-user receiver that performs joint data detection and channel estimation (JDE) of DS-CDMA signals. The proposed sub-optimal receiver is formulated within the framework of the space-alternating generalized expectationmaximization (SAGE) algorithm. The time-varying channel is represented by discrete prolate spherical (DPS) sequences. The resulting receiver iterates between MMSE based channel estimation in the subspace spanned by the DPS sequences and successive interference cancellation at significantly less time complexity than previously proposed SAGE based JDE schemes. The reduction is in the order of 200 for the investigated system parameters. Numerical examples show that the proposed receiver works efficiently in a wide range of velocities and system loads.