Enhanced Spectrum Availability and MU-MIMO Coordination for High Spatial-Spectral Efficiency

  • Maximization of spatial-spectral efficiency, bits per second per Hertz per area (b/sec/Hz/m^2), is a key criteria for wireless system design. Moreover, recently introduced advanced transmission modes such as Multi-User MIMO have yielded order-of-magnitude gains in transmission rate, now realized in both systems and standards. Unfortunately, despite vast gains in raw transmission capabilities, two serious impediments remain to realizing high spatial-spectral efficiency: availability and coordination. Despite wide recognition of the need for improved spectrum allocation, spectrum availability, or legal and regulatory permission to transmit, is today determined at the temporal granularity of no less than 24 hours (white space permission is granted on 24-hour blocks) and with spatial granularity of tens of kilometers. Likewise, once spectrum is determined available, a potentially vast coordination is required to set up the transmission: Which users? Where? When? And using what shared information such as channel state? Moreover, all such coordination itself requires spatial-spectral resources to collect and communicate. In this project, our driving vision is to make new spectrum resources dynamically available and to provide new multi-user coordination mechanisms to enable high-efficiency sharing of all available spectrum.


  • Dynamic Spatial-Spectral Availability with Smart Primary Receivers. The “white space” model of spectrum sharing introduces an important new sharing modality between the extremes of licensed and unlicensed access. In UHF spectrum, it allows frequency bands that are not being used by a TV broadcaster to be repurposed for unlicensed-style access in 24 hour increments. In contrast, we envision a system in which spatial-spectrum availability (or verifiable permission to transmit under regulatory rules) is determined with the aid of smart primary TV receivers rather than by the precalculated range of always-on or always-off primary TV transmitters. Namely, our data obtained from Neilson indicates that even at peak TV viewing times, vast regions that are in range of TV transmitters actually have no receivers on multiple channels. In other words, there is vast spectrum wastage due to having a broadcasting transmitter with sparsely distributed or even no receivers tuned to the channel. We will design, implement, and evaluate a system in which smart TV receivers indicate their channel usage in order to locally and directionally re-use channels that have active broadcasting transmitters, but no locally active receivers. We will show that the secondary users can cancel the well-structured DTV signal and efficiently re-use the spectrum.


  • Foundations for Coordination-Limited Protocols. Once spectrum is deemed available in a region or to a set of users, whether via the aid of smart primaries or via legacy spectrum policies, spectrum users must coordinate spectrum access on a packet-by-packet basis in order to realize high spectral-spatial efficiency during transmission. Despite advances in transmission techniques such as MU-MIMO, our preliminary results indicate that system-wide performance is not dominated by peak physical-layer rates, but rather is fundamentally limited by the coordination required prior to and after a transmission, e.g., sharing of packet-time-scale information such as traffic and channel state. Moreover, such coordination takes place in-band using the same spatial-spectrum resources that could otherwise be used for data. We will design an integrated suite of techniques for coordination-limited protocols that enable highly directional spectrum sharing. We consider the sharing of channel information over users, spectrum, space, and time, to be a foundational resource that must itself be efficiently controlled.


  • At Scale Experimental Platform and Measurement Study. We will design and implement the first open MU-MIMO software-defined radio platform that operates on an order of magnitude frequency range, from 300 MHz to 5.8 GHz. A key component of the platform will be 1 Watt transmit power in order to perform long-range and real-world transmissions including indoor, outdoor, and outdoor-to-indoor transmissions, at operational-network scale. With this unique experimental capability, we will perform a comprehensive set of over-the-air experiments spanning from UHF bands to legacy bands at 2.4 and 5.8 GHz. While prior studies have revealed the basic propagation characteristics across these bands, this will be the first experimental study of wideband MU-MIMO. It will help us understand the fundamental limits of spatial and spectral efficiency that can be achieved with the high directionality of MU-MIMO arrays and flexibility of diverse spectrum bands.