Two rather different forms of wireless communications, one already established and showing dramatic growth and the other only on the horizon, have the potential to impact dramatically the nature of wireless communications in the home and workplace. One is the increasingly high-speed information link typified by 802.11x/802.15, first applied to broadband internet access but moving toward video and audio entertainment distribution as well, and the other is the sensor network, which can offer security, medical monitoring, and a variety of other important services. Unfortunately these usages are largely incompatible. High-rate applications will generally have adequate power sources for both transmission and processing purposes. Service interruptions have the potential to be annoying, but they are not, in general, threatening to health or safety. Sensor networks, on the other hand, involve messages which are generally short and infrequent, but which may have a high level of importance (consider intrusion or fire alarms, or child
monitors). Moreover, such sensors (and less critical environmental sensors) often require placement that is independent of permanent power sources, and thus require energy efficient operation.
Normally, such disparate requirements would suggest the separation of these applications into different frequency bands, but the emergence of unlicensed bands suggests that designers will not have the option of such comfortable isolation. It is therefore important to consider how such applications may coexist within the same frequency space. In particular, one may speculate that the potential for such coexistence would be enhanced within the private space that we find in homes and businesses - a volume in which radio transmission may be beneficially controlled by a single entity, and in which intra-system interference dominates. It is the premise of this proposal that solutions exist that can allow a wide variety of disparate applications to coexist efficiently within such a constrained, but controlled space. In particular, the researchers believe that coordinated usage characterized by multiple antennas and multiple appliances (MAMA) represents a new type of network, and offers significant opportunities to interwork such disparate systems efficiently.
To accommodate this mixed set of bit-rate and energy requirements, the researchers propose a comprehensive approach that encompasses link layer, MAC layer and cross-layer techniques. At the link layer, the research proposes:
-Flexible bandwidth modulation formats and spatial multiplexing and diversity. With the device transmission rate set by the application, the signaling bandwidth of a device is optimized versus a power level determined either by interference avoidance considerations or by energy constraints. We will argue that this leads to ultra-wideband (UWB) modulation characterized by low-signal-to-noise-ratio (low-SNR).
-Multiple transmit and receive antennas to be utilized in several modes: (1) in multi-input multi-output (MIMO) mode, the channel is harnessed to increase bit-rate; (2) in diversity mode, spatial diversity enhances power efficiency; (3) in beamforming mode, transmitter antennas direct energy away from other devices and networks. Consistent with UWB signaling, our research will focus on the low-SNR regime for MIMO systems.
Since link layer techniques cannot address all the complex requirements of MAMA networks, the researchers propose a cross-layer approach that takes into account the distributed, non-cooperative nature of the networks to achieve a more efficient use of the network power and bandwidth resources. The cross-layer approach consists of two components:
-Game theoretic methods: based on utility functions that depend on transmitted power and throughput, users
adapt their transmitters and receivers to maximize their individual utilities.
-'Thin' MAC layer protocols: for given physical layer transmitters and receivers, incremental redundancy hybrid ARQ provides non-collaborative rate adaptation and spectrum resource sharing.
The proposed space-frequency-cross-layer approach to designing wireless networks is distinctly different from conventional network design where data rate is maximized for a given bandwidth and power, generally ignoring intra- and inter-network interference. The juxtaposition of UWB, MIMO and cross-layer techniques in MAMA networks spans a multidimensional signal space that will create a rich set of research problems and network architectures.
Broader Impact: The work proposed here (if successful) will make possible the coexistence of new high-speed
wireless applications with emerging sensor networks at home and in the workplace. Although the discussion focuses on the home and work environments, it should be understood that the applications are more widespread - to hospitals, factories, and some robotic scenarios. MAMA networks also provide an exciting platform for the educational goals of the academic institutions, including activities for both undergraduates and graduate students.
The research institutions involved in this proposal are Princeton University, the New Jersey Institute of Technology, and Rutgers University, and The Wireless Communications Dept., Bell Labs, Lucent. The work will be done under the auspices of the N.J. Center for Wireless Telecommunications (NJCWT). The NJCWT is an inter-institutional research and educational organization sponsored and funded by the N.J. Commission on Science and Technology.
|Effective start/end date||1/1/04 → 12/31/07|
- National Science Foundation: $627,864.00