Noncoherent eigenbeamforming and interference suppression for outdoor OFDM systems

We investigate a new approach to uplink communications in wideband outdoor cellular systems that can take advantage of multiple antennas at the base station in a scalable manner, while eliminating or minimizing overhead for channel estimation. The proposed techniques, which focus on exploiting correlated channels with the use of closely spaced antenna arrays, are applicable to emerging Orthogonal Frequency Division Multiplexing (OFDM) based Wireless Metropolitan Area Network (WMAN) systems, such as those based on the IEEE 802.16/20 standards. Outdoor channels frequently have a small number of dominant spatial modes, which can be learned from overhead-free estimation of the spatial covariance matrix by averaging across subcarriers. We describe an eigenbeamforming receiver which projects the received signal along the dominant spatial modes, yielding a beamforming gain that scales up with the number of receive elements and a diversity level depending on the number of dominant spatial modes. Shannon limits are first computed for block fading approximations to time- and frequency-selective channels. The suboptimal noncoherent diversity-combining receiver is shown to approach these limits, with linear complexity in the number dominant modes. Further, for dealing with spatially non-white interfering signals, adaptive suppression techniques are shown to mitigate strong interference with minimal training overhead.