Blind MMSE-ZF receiver and channel identification for asynchronous CDMA in multipath channels

We consider an asynchronous DS-CDMA system operating in a multipath environment. The received cyclostationary spread signal is oversampled with respect to the chip rate (and/or is received at multisensors) and is converted to a stationary vector signal, leading to a linear multichannel model. The blind MMSE-ZF receiver is derived and an estimate of the channel is obtained using the spreading sequence properties and second-order statistics. The receiver turns out to be the exact extension to the general asynchronous multipath case of the constrained minimum-output energy (MOE) receiver originally proposed in [1]. Dimensional requirements for the determination of the receiver and channel identification are derived and extension to the case of sparse channels is made. It is observed that especially for the sparse scenario, an estimate of the desired user channel is obtained in a small number of data samples, despite the fact that the processing gain is quite large and that severe near-far conditions are prevalent. This makes the algorithm particularly suitable for the case of slow fading in the UMTS WCMDA, where, multiuser algorithms can be easily implemented once accurate channel estimates are available.