Andrea Goldsmith

Fundamental Capacity Limits and Cross-layer Design in MIMO Systems

Wireless systems of the future must support ubiquitous multimedia communications between people as well as devices. There are many research challenges associated with such systems, including limited bandwidth, random variations in the wireless channel, and battery constrants in small radio transceivers. Many of these challenges can be overcome with multiple antennas at the transmitters and receivers of the wireless network, since these antennas both increase data rate and reduce channel randomness. In this talk we will first explore the fundamental capacity limits and capacity achieving strategies of wireless channels and networks with multiple antennas. Of particular interest are asymptotic results on optimality of beamforming for multiuser systems as well as the capacity of virtual MIMO systems. Then we will investigate multiple antennas and more general MIMO networks in the context of cross-layer design. MIMO systems have inherent tradeoffs between multiplexing and diversity that extend across multiple layers of the network protocol stack. Thus, the tradeoffs inherent to these techniques must be optimized across protocol layers. We will discuss optimizing these inherent tradeoffs in both source/channel coding and energy-constrained networks. Performance results will be presented that demonstrate significant improvements through cross-layer design.