Blind Channel Estimation and Multiuser Detection for Multi-Rate CDMA Communications

The future wireless communications systems should be able to offer wide variety of applications, which have vastly different quality of service (QoS) requirements. The time-variable QoS may require the support of variable bit rates on the wireless links to the individual users. Multi-rate DS/CDMA is a promising basis on which to support the variable bit rates on the individual wireless links. Currently, the study on channel estimation and multiuser detection for multi-rate DS/CDMA, which makes full use of the nature of multi-rate signals, is still at its early stage. The thesis deals with the application of subspace-based techniques to blind channel estimation and multiuser detection for multi-rate DS/CDMA, including single-carrier and multicarrier scenarios. For the single-carrier case, space-time blind linear multiuser detection is investigated for synchronous dual-rate systems over the AWGN channel. The performance is evaluated analytically. The multi-rate generalization and the asynchronous extension are discussed. Two-stage space-time dual-rate blind detectors are also presented. Furthermore, blind adaptive channel estimation and detection schemes for asynchronous dual-rate systems over frequency-selective multipath channels are developed. In the context of multicarrier DS/CDMA, based on a finite-length truncation approximation on the band-limited chip waveform, blind timing acquisition and channel estimation scheme is proposed for multi-rate systems. The channel estimation error due to the finite-length chip waveform truncation is analyzed by exploiting a first-order perturbation approximation