|You are kindly invited to:|
|Communication and Information Theory Talk|
Dr. Thomas L. Marzetta
Bell Labs, Alcatel-Lucent
On: The Analytical Design of Massive MIMO Systems
Massive MIMO entails a numerically large base station array, say 128 elements or more, serving a smaller number of single-antenna terminals, typically 16 or more, using the same time/frequency resources. This aggressive multiplexing confers attractive spectral efficiency gains over 4G technology. The all-important acquisition of channel-state information by the base station is derived from uplink pilots and time-division duplex reciprocity. The ability to focus power selectively on the downlink, and selectively to collect power on the uplink, enables Massive MIMO to reduce radiated power drastically. Power control really works, providing an unmatched user experience even at the edges of the cell. The combination of increased throughput and reduced radiated power implies superior energy efficiency. The healthy excess of service-antennas over terminals permits simplicity in Massive MIMO
operation: linear precoding/decoding operators are competitive with dirty-paper coding and exhaustive multiuser detection, and optimal power control laws require only linear programming. An end-to-end simulation of a Massive MIMO cellular system would be a daunting task, typically involving a total of 6000 or more service-antennas and 800 or so terminals. Paradoxically the large numbers of service antennas make performance analyses much easier than one could reasonably anticipate.
Bayesian techniques yield exact non-asymptotic expressions for received signal-to-interference ratios which translate directly into lower bounds on instantaneous capacity. The SINR expressions account for receiver noise, channel estimation error, channel non-orthogonality, the detailed linear precoding/decoding that is used, power control, non-coherent inter-cell interference, as well as coherent inter-cell interference arising from pilot contamination.
Thomas L. Marzetta Marzetta was born in Washington, DC, and he received the PhD in Electrical Engineering from MIT under supervision of Prof.
Arthur B. Baggeroer. After careers in petroleum exploration and defense he joined Bell Labs in 1995, eventually becoming director of the Communications and Statistical Sciences Research Department. He is the originator of Massive MIMO, and he currently heads the Large-Scale Antenna Systems Project. He was the recipient of the IEEE 1981 ASSP Paper Award from the Signal Processing Society, became an IEEE Fellow in 2003, received the 2013 IEEE Guglielmo Marconi Best Paper Award, and, with Hong Yang, he received the 2013 IEEE OnlineGreenComm Best Paper Award.