People
Prof. Heather Zheng
Ashwin Sampath (wireless routing)
In the News
Prof. Heather Zheng selected as TR35: 2005 MIT Technology Review Top 35 Innovators under the
age of 35 (UCSB & UMCP
news)
Prof. Zheng's talk at UC Berkeley Networking Seminar on Managing Open Spectrum Systems , Oct. 10,
05; also at Bell-Labs, NJ, Sept. 05.
2 papers accepted by Dyspan05 (one to be presented at the plenary
session)
CS290F in Winter 2006 on intelligent
wireless systems
Links
Prof. Ben Zhao &
UCSB Current Lab
Conferences
Dyspan05 11/8-11/11
mobihoc06 CFP 12/2
Contact
Prof. Haitao (Heather) Zheng
Engineering I, 2159
Computer Science, UCSB
Santa Barbara, California 93106-5110
Phone: +1 (805) 893-3560
Fax: +1 (805) 893-8553
Intelligent Networking
The fundamental concept behind Intelligent Networking is to push Human
brain-power into end-devices. Human are intelligent and adaptive. We sense neighborhood
environment, adapt to dynamics and variations, and learn from past
experiences to change future behavior. This is so called cognitive cycle.
Our approach is to apply the concept of cognitive to network elements that allows networks
to manage themselves in a self-aware and adapative manner. Areas to be addressed include
self-organizing networks, adaptive routing and MAC protocol design, network resource
management, topology discovery/control, security as well as network
infrastructures.
- Application of reasoning and learning to facilitate improved efficiency, performance, fault-tolerance, security and other funcationality of networks.
- Introduction of adaptive self-organization to provide fault-tolerant, reliable operations in resource constrained networks.
- Introduction of device collaboration to optimize network operation in a distributed way
- DART (Adaptive Routing) - reliable and resource efficient routing protocols
- SMART (Spectrum Aware Routing) - exploiting interactions between MAC spectrum management and Routing
Associated Projects:
Fundations
Our projects are built on top of prior research experiences on
cognitive radio and open spectrum systems. Prior work on MAC layer
issues such as dynamic spectrum allocation and distributed coordination provides initial
building blocks towards the implementation and deployment of a fully
operational cognitive networks. We are currently addressing networking
and MAC level issues, which also extend the concept of cognitive radio into the
general aspects of system and networking.
Previous work focuses on the design of decentralized infrastructure
for Open Spectrum system, where devices interact to efficiently utilize
spectrum, while optimizing for traditional goals such as fairness and
maximizing utilization. The project combined expertise from multiple areas:
distributed artificial intelligence, game theory and wireless communications
& networking together. This work has produced efficient distributed
algorithms for spectrum management and coordination, backed up by strong
theoretical results.
Associated Projects:
-
Collaboration and Fairness in Spectrum
Access
We show that collaboration among devices, together with appropriate regulations, provides a strong base for fair and efficient spectrum usage. When direct collaboration is not possible (due to signaling compatibility, radio design, protocol constraints, etc.), rule based spectrum management where each device follows a network-enforced rule set would provide similar performance as in collaboration based approaches. These results make Open Spectrum communication efficient and cost-effective, paving the way to large scale deployment. . - Distributed Coordination in Dynamic Spectrum
Environments
Collaboration among devices requires efficient coordination. In addition, traffic coordination between devices are required for reliable communications. That is, each transmitter and receiver pair needs to synchronize their spectrum usage. Coordination requires a signaling path among devices for information exchange. Unlike fixed spectrum systems, spectrum heterogeneity in open spectrum systems prevents the use of a common coordination channel. Instead, secondary devices must select them dynamically, while minimizing the total number of coordination channels in the network and maintaining connectivity. We propose a distributed, scalable and efficient coordination framework where devices dynamically select their coordination channel based on local conditions. Without relying on the existence of a pre-assigned common control channel, HD-MAC carries potential to provide robust operation under network dynamics. While this approach can be implemented by upgrading the legacy protocol stack without modifying the MAC protocol, modifications to the MAC protocol that address spectrum heterogeneity and significantly improve system performance.
A Short Introduction to Cognitive Radio and Open Spectrum
Wireless devices are becoming ubiquitous, placing
increasing stress on the fixed radio spectrum available to all access technologies. This
leads to spectrum scarcity problem. To eliminate interference between different wireless technologies,
traditional (and current) policies allocate a fixed spectrum slice to
each technology, i.e. command and
control. This static assignment prevents devices from efficiently
utilizing allocated spectrum, resulting in spectrum holes (no targeted
devices in local area) and very poor utilization (6-10%) in other
geographic
areas (source DARPA). Given that the current spectrum licensing policy
facing near-future threat of spectrum scarcity and the increasing crowd
in unlicensed spectrum band, efficient spectrum management is necessary
and critical to future development of system and networking.
Enabled by software defined radio (SDR)
technology, Open Spectrum allows unlicensed (secondary) users to share
spectrum with legacy (primary)
spectrum users, thereby "creating" new capacity and commercial
value from existing spectrum ranges. Based on agreements and
constraints
imposed by primary users, secondary users opportunistically utilize
unused
licensed spectrum on a non-interfering or leasing basis. Open Spectrum
systems have a high potential for impact, and can enable large segments
of the world population to be connected efficiently and cost
effectively using a variety of devices, while making world-wide
deployment easy for service providers. It is envisioned to have an
impact on wireless system and networking like the way that Internet and
packet switching have reshaped the world.
Open spectrum requires a new line of
cognitive radios. Cognitive refers to a device's ability to sense
surrounding environment conditions and adapt its behavior accordingly.
This process is often referred to as cognitive cycle: (sense -
characterize - learn - adapt). The advent of
cognitive radio clearly portends a potential revolution in
wireless networking. It makes every aspect of wireless transmission and
reception
programmable. Devices can intelligently select the transmission format
and
media access technology, while all these characteristics are
software-reconfigurable. Most importantly,
devices equipped with cognitive radio can constantly monitor the
spectrum and discover how the
spectrum is being used. By dynamically
reconfiguring it to transmit in a way that does not interference with
existing
users, cognitive software radio allows exploitation the portion of
spectrum
that is allocated but unused. This creates additional capacity and
revenue from
existing spectrum, and addressing spectrum scarcity problem. The
concept of cognitive radio has received enormous attention from both
government (FCC, DARPA, NSF, ARMY, ONR) and major industrial
sources. In addition, the introduction of cognitive radio could
potentially restructure the industry, significantly increasing
the role of software in devices, and the role of end-devices in system
and networking.
