Blind Adaptive Space-Time Processing for Cyclostationary Signals

In this paper we present a blind adaptive space-time processing algorithm for separating signals which are spectrally and/or spatially overlapped. The algorithm exploits the cyclostationary nature of many communication signals, but does not require knowledge of the statistical properties of the desired signal. It merely requires knowledge of a (distinct) cycle frequency. It is shown that the performance of the algorithm converges at a rate O(1/N), where N is the number of received samples, to the performance of the optimal (trained) receiver with the given structure. Furthermore, we provide an (algebraic) analysis of the performance of a multiuser communication system which employs our receiver, and confirm this result in simulation examples. These examples demonstrate that our cyclic adaptive space-time processor is an attractive alternative to beamforming or filtering alone. ∗Corresponding author. Department of Electrical and Computer Engineering, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada, L8S 4K1. Telephone: +1-905-525-9140, Ext. 24098. Fax: +1-905-521-2922. Email: [email protected]