Existing 802.11n is inefficient whenever nodes with a different number of antennas share the medium. This research aims to develop an advanced 802.11n design that delivers as many concurrent transmissions as permitted by the MIMO transmitter with the maximum number of antennas. Such a design delivers performance gain while maintaining the fully distributed random access nature of today's 802.11n, i.e., its ability to support bursty traffic, distributed decisions, and ad hoc deployment. The research develops a novel carrier sense mechanism that enables MIMO nodes to detect whether they can transmit in the presence of an ongoing transmission without interfering with it. It then builds on this mechanism to deliver a random access protocol where MIMO nodes contend for both time and degrees of freedom, without any form of centralized coordination. Finally, it implements its design and evaluates it in a wireless testbed. By delivering a higher throughput for the same resource, this research advances the design of MIMO 802.11 networks and the corresponding industry.

Project Report

Wireless interference is perhaps the most fundamental limitation of today’s WiFi and cellular networks. It is the main reason why today’s network cannot deliver higher data rates and suffer from congestion in the presence of many users. The goal of this project is to combat wireless interference using MIMO technologies. In particular, the research deliver three types of technologies: First, the research develops MegaMIMO, a clean-slate design for Wi-Fi transmitters. MegaMIMO transmitters carefully construct wireless signals so that interfering signals will sum up to zero at the receivers, while desired signals persist. In other words, Wi-Fi access points can simultaneously transmit to their respective users, without any interference. Hence for the first time, wireless networks can scale linearly with the number of users, while re-using the same amount of spectrum. A prototype implementation of MegaMIMO has demonstrated 10x gain in throughput over 802.11n Wi-Fi. Second, deploying MegaMIMO will need long-term investment in new hardware and standards. Thus the research also investigates whether one can combat interference with today's Wi-Fi devices. The research develops OpenRF, the first purely software system to bring MIMO interference cancellation techniques to commercial Wi-Fi devices. OpenRF showed that such techniques bring massive gains to today's mobile applications for the first time. Third, the research also explored whether motion can help in combatting interference. It introduced, MoMIMO, a new antenna design that slides by just an inch or less on a Wi-Fi access point to dramatically reduce interference. At the heart of this design are new wireless models that understand how signals add up differently across spatial locations. Such antennas can easily be connected to the external ports of today's Wi-Fi access points to reduce interference. Further, one can leverage the natural mobility of handheld devices themselves, advising users on how best to move them, when they experience interference.

Agency
National Science Foundation (NSF)
Institute
Division of Computer and Network Systems (CNS)
Type
Standard Grant (Standard)
Application #
1117194
Program Officer
Thyagarajan Nandagopal
Project Start
Project End
Budget Start
2011-09-01
Budget End
2014-08-31
Support Year
Fiscal Year
2011
Total Cost
$300,000
Indirect Cost
Name
Massachusetts Institute of Technology
Department
Type
DUNS #
City
Cambridge
State
MA
Country
United States
Zip Code
02139