An integrated signal processing framework for multiuser CDMA communications

The major signal processing challenges in wireless CDMA systems stem from time-varying multipath propagation effects, multi-access interference (MAI), and the complexity of the DSP algorithms. We propose a framework based on canonical multipath-Doppler coordinates for addressing these issues in an integrated fashion. The canonical coordinates are derived from a fundamental characterization of channel propagation dynamics in terms of discrete multipath-delayed and Doppler-shifted copies of the transmitted waveform. These delayed and Doppler-shifted spreading waveforms constitute a natural fixed basis and dictate a canonical low-dimensional processing. In addition to providing a robust vehicle for channel modeling and estimation the framework facilitates exploitation of dispersion effects for MAI suppression via subspace-based processing in the canonical coordinates. Finally, the low-dimensional subspace formulation affords a direct handle on the complexity of the DSP algorithms