Channel estimation and multiuser detection in long-code DS/CDMA systems

The problems of channel estimation and multiuser detection for direct sequence code division multiple access (DS/CDMA) systems employing long spreading codes are considered. With regard to channel estimation, several procedures are proposed based on the least-squares approach, relying on the transmission of known training symbols but not requiring any timing synchronization. In particular, algorithms suited for the forward and reverse links of a single-rate DS/CDMA cellular system are developed, and the case of a muitirate/muiticode system, wherein high-rate users are split into multiple virtual low-rate users, is also considered. All of the proposed procedures are recursively implementable with a computational complexity that is quadratic in the processing gain. With regard to the issue of multiuser detection, an adaptive serial interference cancellation (SIC) receiver is considered, where the adaptivity stems from the fact that it is built upon the channel estimates provided by the estimation algorithm . Simulation results show that coupling the proposed estimation algorithms with a SIC receiver may yield, with a much lower computational complexity, performance levels close to those of the ideal linear minimum mean square error (MMSE) receiver, which assumes perfect knowledge of the channels for all of the users and which (in a long-code scenario) has a computational complexity per symbol interval proportional to the third power of the processing gain.