Sparse equalization in high data rate WCDMA systems

Low computational complexity is among the most important criteria for receiver design in mobile communication systems such as WCDMA. Low complexity receivers are desirable especially in the downlink. In this paper we address the problem of complexity reduction of MMSE equalization. It is well known that for good performance the length of an equalizer should be at least twice the channel length. However, employing such a long filter is computationally very demanding. A novel method for selecting the filter coefficients for WCDMA receiver is proposed. This receiver is called sparse equalizer. It is shown via simulations to outperform both conventional RAKE and generalized RAKE (CRAKE) using the same filter length. Simulations are carried out in high speed downlink packet access (HSDPA) systems in ITU vehicular A and pedestrian A channels with multiple receive antennas.