Faculty of Engineering and Natural Sciences
Link Adaptation in MIMO Interference Systems
Dr. M. Fatih Demirkol
Hawaii Center for Advanced Communications
University of Hawaii at Manoa
In MIMO interference systems, multiple links, each with different transmitter-receiver pairs, are allowed to transmit in a given range possibly through multiple streams per link. Such a network of interfering MIMO links can yield high spectral efficiency with optimized signaling that uses the spatial processing capabilities at both ends of the links according to the channel and interference conditions. Adaptive beamforming done at the transmitter for each stream using the knowledge of the channel state information (CSI) not only affects the efficiency in directing the transmitted power towards the designated receiver but also shapes the interference emitted to the other links. Accordingly, the capacity of a link is dependent on all transmissions in the range.
The need for optimizing the link parameters with local information and reasonable computational burden for such a system motivates a decentralized approach. The set of parameters for each link includes the number of streams to be transmitted, array weights for each stream and power distribution among the streams.
In this talk, we take a look at the link adaptation and stream control problem in such a network from a game-theoretical point of view. We introduce a two-level hierarchical approach that uses coordination among links based on a variant of a learning algorithm called “adaptive play” for stream control at the higher level. At the lower level of the hierarchy, however, each link selfishly pursues the maximization of its own mutual information subject to the limit on the number of data streams to which it agreed at the higher level.
June 26, 2006, 13:40, FENS G032