Robust multiuser detection using Kalman filter and windowed projection approximation subspace tracking algorithm

"The authors propose some robust adaptive multiuser detection schemes for direct-sequence code-division multipleaccess multipath frequency-selective fading channels. Multiple access interference (MAI) and intersymbol interference (ISI) are presented in an identical format using expanded signal subspace, which facilitates multiuser detection in a symbol-bysymbol fashion. This study contributes to the theoretical aspect of adaptive multiuser detection by proving that the optimum linear multiuser detectors that achieve maximum signal-to-interference-plus-noise ratio (SINR) must exist in the signal subspace, and the theoretic SINR upper bound is also derived. Another contribution of this study is to propose the design of multiuser detectors in an expanded signal subspace, and introduce subspace estimation and Kalman filtering algorithms for their adaptive implementation. To robustify the adaptive detectors against subspace estimation and channel estimation errors, a modified projection approximation subspace tracking (PAST) algorithm is proposed for subspace tracking. It is demonstrated by simulations that these adaptive detectors effectively suppress both MAI and ISI and converge to the optimum SINR. They are robust against subspace estimation errors and channel estimation errors compared to the conventional Wiener minimum mean square error (MMSE) detector.