MMSE joint detection in frequency-selective wireless communication channels for DS-CDMA systems

In digital mobile communication systems, two important problems arise, namely multiple access interference (MAI) and intersymbol interference (ISI). An attractive solution is the use of code division multiple access (CDMA) systems possibly combined with TDMA and FDMA. In CDMA communication systems, the MAI and ISI can be efficiently combated by applying multi-sensor linear joint detection (JD) techniques, in particular, space-time processing methods. In most cases, they are derived from a symbol-level error criterion. However, in some other cases they are obtained from a chip-level viewpoint, also known as chip sequence restoration techniques. In the present paper, a detailed analytical comparison between chip-rate and symbol-rate minimum mean square error (MMSE) block filtering is provided, taking into account also other features such as structured vs. non-structured block filter computation and chip-level vs. symbol-level training sequence. Monte-Carlo simulation results are also given to provide realistic comparisons of the different methods.