Outline Perspectives and Challenges for Cognitive Radio Networks Location Awareness for Dynamic Spectrum

Outline
‰ Background
Perspectives and Challenges for
Cognitive Radio Networks
‰ What is Cognitive Radio (spectrum overlay
technique)?
‰ Challenges in Cognitive Radio
‰ Location Awareness for Dynamic Spectrum
Access in CR networks
Li-Chun Wang (王蒞君)
‰ Conclusions
Department of Communications Engineering
National Chiao Tung University
Hsinchu,Taiwan
lichun@cc.nctu.edu.tw
3C Tradeoff in Wireless
‰ The 3C (Capacity, Cost, Cwality)
Tradeoff
High Capacity
Good Quality
Low Cost
‰ The more things changes, the more
remain the same (Alphonse Karr, 1808)
What is the fundamental issue
for wireless communications?
Spectrum Efficiency
RF Spectrum – 1.39 to 5.923 GHz
Motivation
‰ Licenced TV band Usage: Over the
air/(Cable + DBS) is 15% to 85%
‰ Cognitive Radios (CRs) can
Most (85%-90%) of the
spectrum is unused.
Mid-Band
opportunistically use spectrum
white space and increase usage by
10x
Evidence of overcrowding
High-Band
Evolution of Wireless Communications
Both licensed/unlicensed
band
Licensed band
IEEE 802 LAN/MAN/RAN Standards
‰ WLAN (IEEE 802.11)
¾ WiFi 802.11a/b/g/n
‰ WPAN (IEEE 802.15)
¾ Bluetooth 802.15.1
¾ Co-existence 802.15.2
¾ High-data-rate UWB 802.15.3a
¾ High-data-rate UWB 802.15.4a (Zigbee)
‰ WMAN (IEEE 802.16)
‰ WRAN (IEEE 802.22)
Unlicensed band
Current Situation of Spectrum Usage
‰ Allocated spectrum does not mean
licensed.
‰ Licensed does not mean used.
¾
佔著____,
不_____。
Future Spectrum Sharing
‰ New unlicensed bands
¾ 60 GHz
‰ Spectrum underlay
¾ Ultra-wideband radios
¾ 802.15.3a, 802.15.4a
‰ Spectrum overlay
¾ Cognitive radios
Basic Idea of UWB
Basic Idea of Cognitive Radio
Cognitive frequency-agile radios can be allowed to
intelligently “switch lanes” and adjust parameters.
Source: Fantasma
Outline
Why Cognitive Radio?
‰ Pros and Cons for Cognitive Radio:
¾解
決佔
著____,
不
_____的
問
題
。
¾ 你幹
嘛
管
我
家
廁
所
有
沒
有
人
用?
‰ The true is
¾ Current unlicensed bands are too crowded.
¾ Emergency application
¾ Provide ubiquitous wireless services by
heterogeneous wireless networks
What is Cognitive Radio?
‰ Not CR in your keyboard!
‰ Co-exists with legacy wireless systems
‰ Utilize the spectrum resource of the
legacy system.
‰ Does not cause interfere to them
‰ Background
‰ What is Cognitive Radio (spectrum overlay
technique)?
‰ Challenges in Cognitive Radio
‰ Location Awareness for Dynamic Spectrum
Access in CR networks
‰ Conclusions
Definition
‰ Cognitive Radio (CR) is a radio that can
change its parameters based on
interaction with the environment in
which it operates.
Application Scenario 1
‰ The IEEE 802.22 Wireless Regional Area
Network (WRAN)
¾
¾
Fixed point-to-multipoint wireless regional
area networks for wireless broadband
access (especially in rural areas)
Reuse TV broadcast bands on a noninterfering basis
High power CR device vs Low power CR device
‰ High power “fixed/access" unlicensed devices
¾
Incorporate a geo-location method such as GPS or be
professionally installed to determine their geographic
coordinates, which would be used in conjunction with a
database to identify vacant channels.
¾
1 W peak transmit power, 4 W EIRP (6 dBi antenna)
‰ Low power “personal/portable” unlicensed devices
¾
Operate only when they receive a control signal from a
source such as an FM or TV station that identifies the
vacant TV channels in that particular area
¾
100 mW peak transmitter power
‰ FCC Docket No. 04-186 / 06-156
Application Scenario 2
‰ Super Wi-Fi network (urban areas)
An informal coalition of technology companies,
including Dell, Google, Hewlett-Packard, Intel,
Microsoft, and Philips Electronics, has been formed.
¾ Ask the FCC to make unused portions of the TV
broadcast spectrum available for unlicensed use by
wireless devices.
¾ A significant portion of the TV signal spectrum should
become available when TV broadcasters move from
analog to digital broadcasting
¾
Fundamental Goals
‰ Primary protection
¾ Theoretically provable and practically
feasible to protect primary as both fixed
devices and portable/personal devices
‰ Secondary coexistence/QoS support
¾ Technical challenges exist and careful design
needed
Key goals of a cognitive radio
Design elements of a cognitive radio network
DSA
‰ Dynamic Spectrum Access (DSA): quickly and
Secondary
coexistence/QoS
Primary
protection
robustly detect the presence of incumbent users,
who have preemptive rights to access the
spectrum, to avoid causing interference to these
users,
‰ Dynamic Spectrum Sharing (DSS): CRs must be
aware of other cognitive radio networks of likely
similar access rights and coexist with these
networks, and,
‰ Dynamic Spectrum Multi-channel operation
(DSM): CRs must be spectrum-agile and provide
seamless operation across multiple channels,
potentially simultaneously.
Application
User utility/Policy
QoS requirements
Spectrum handoff
Delay/Jitter/Loss
Inter-system and
Delay/Jitter/Loss
inter-flow
communication
Spectrum Handoff
Load (node &
Interference aware
radio resource)
routing
balancing
Transport
Network
LLC
Spectrum aware
Policy
routing
Delay tolerant
Topology
networking
management for
(DARPA DTN)
distributed sensing
Radio
Centralized (e.g.,
environment
broker) or distributed
characterization (e.g., opportunistic)
Policy
Power control
Interference temperature
Dynamic frequency
selection (if channel
vacation)
Link
Coordination of quiet
periods
Directionality
Spectrum sensing
management
MAC
Physical
DSS
QoS requirements
User utility/policy
Policy
Interference aware
dynamic frequency
selection
Power control
Channel assignment
Centralized
(e.g., broker) or
distributed
(e.g.,
opportunistic)
Space, time, and
code division
multiple access
Fairness
Spectrum sensing
algorithms
Wideband or
narrowband sensing
Low SNR signal
detection
Multiple antenna
beamforming and
beamnulling
Adaptive
modulation and
coding
Waveform shaping
Two Main Characteristics for CR
‰ Cognitive capability:
¾ The ability of the radio to capture or sense
the information from its radio environment
¾ Monitoring the power in some frequency
band is NOT enough
¾ Sophisticated techniques are needed to
capture the temporal and spatial variations
in the radio environment
‰ Re-configurability:
¾ Enable the radio to be dynamically
programmed
Multi-channel
assignment
Multi-channel
and multi-path
routing
Multi-channel assignment
(Non)-Coordinated
resource sharing
(Non-)
contiguous multichannel operation
Inter-channel
synchronization
Directionality
Real-time and
dynamic resource
allocation
Multi-channel
access
Multiple antenna
beamforming and
beamnulling
Spreading
Wideband
antennas
ADC
Multicarrier
modulation
Programmable
filters
Multiple
antennas
Beacon detection
DSA
DSM
QoS
Multi-layer
requirements
management
Delay/Jitter/Loss Multi-flow
management
DSS
Bandwidth (RF,
BB) scalability
DSM
Key Functions for CR
‰ Sensing Radio
Wideband antenna, PA, and LNA
¾ High speed A/D & D/A
¾ Scalable for MIMO
¾
‰ PHY layer
OFDM transmission?
¾ Dynamic frequency selection, adaptive modulation
and coding, power control
¾
‰ MAC layer
Transmission parameter optimization
Rate adaptation
¾ Feedback mechanism
¾ Negotiate and opportunistically use of radio resource
¾
¾
Key Properties for CR
‰ RF technology to listen to huge spectrum.
CR Network Architecture
‰ On licensed band
‰ Knowledge of primary users’ spectrum
usage in terms of location and time.
‰ Rules of sharing the available resource
(time, frequency, location)
‰ Intelligence to determine the optimal
transmission parameters (bandwidth,
power, QoS).
Outline
CR Network Architecture (Cont’d)
‰ On unlicensed band
‰ What is Cognitive Radio (spectrum overlay
technique)?
‰ Challenges in Cognitive Radio
‰ Location Awareness for Dynamic Spectrum
Access in CR networks
‰ Conclusions
Key Challenges
‰ Physical Architecture:
¾ An accurate detection of weak signals of
licensed users over a wide spectrum range.
¾ Implementation of RF wideband from-end
and A/D converter (multi-GHz speed)
Key Challenges (Cont’d)
‰ Re-configurability:
¾ Adjusting operating parameters:
Operating frequency
Modulation
9 Transmission power
9 Communication technology
9
9
Challenges are to reconfigure the
transmission parameters not only
at the beginning of a transmission
but also during the transmission
Key Challenges (Cont’d)
‰ Cognitive capability:
Spectrum sensing (detecting spectrum holes)
¾ Spectrum analysis/decision (determining data rates,
transmission rate, and modes)
¾ Spectrum mobility (if the current spectrum band in
use become unavailable)
¾ Spectrum sharing (providing fair spectrum scheduling
among coexisting CR users)
¾
Challenges lies in the fact that radio environment
changes over time and space.
Concurrent transmission is an ambitious but
promising goal.
Research Topics in Cognitive Radio
‰ Spectrum Sensing
‰ Spectrum Decision
‰ Spectrum Mobility
‰ Spectrum Sharing
Research Topics in
Spectrum Sensing
Research Topics in
Spectrum Decision
‰ Non-cooperative transmitter detection
‰ Cooperative transmitter detection
‰ Interference-based detection
‰ Challenges:
¾ Wideband sensing and multi-band agile
wireless networks
¾ Spectrum sensing in multi-user network
¾ Interference temperature measurement and
metrics
¾ Fast Detection capability
Research Topics in
Spectrum Mobility
‰ QoS provisioning in adaptive, spectrum,
and cognitive network
‰ Spectrum mobility in time (channel
variation)
‰ Spectrum mobility in space (user
movement)
‰ Spectrum handoff
‰ Multi-parameter decision model for
spectrum analysis
¾
Only SNR is not enough
‰ Multi-band spectrum decision
‰ Spectrum decision over heterogeneous
licensed/unlicensed spectrum bands
Research Topics in Spectrum Sharing
‰ Overlay spectrum sharing vs underlay
spectrum sharing
‰ Cooperative spectrum sharing vs noncooperative spectrum sharing
‰ Centralized spectrum sharing vs
distributed spectrum sharing
‰ Inter-network spectrum sharing
Why Is Location Awareness
important for Cognitive Radio?
Outline
‰ Background
‰ What is Cognitive Radio (spectrum overlay
technique)?
‰ Challenges in Cognitive Radio
‰ Location Awareness for Dynamic Spectrum
Access in CR networks
‰ Conclusions
Location Awareness and Spectrum
Sensing
‰ Wide spectrum sensing, identification and
spectrum opportunity utilization are the three
important techniques for CR systems.
‰ Wide spectrum sensing is not free:
¾ Energy consumption
¾ Access delay
‰ Issue:
¾ Should a CR device require to sense the spectrum
anywhere and anytime?
‰ Location awareness is needed for compliance
¾
to conform to the rules of the country where it
operates.
‰ Location awareness is needed for efficiency.
¾ to
verify distance between devices
¾ to enable use of higher beacon power
‰ A Dynamic Spectrum Access system must have
location awareness in its mobile devices [John
Chapin’06]
Objective
‰ Dimension the region where a CR device
can reuse the same frequency band of
the primary user
¾ without the need of scanning the
whole spectrum
¾ by taking advantage of the location
information.
Problem Formulation
‰ The locations of the infrastructure user MS3 (r3,θ3) the
receiver of the ad hoc link MS2 (r2,θ2), and BS(0,0)
are fixed.
‰ MS1 (r1,θ1) is uniformly distributed in the coverage of
the BS, πR2.
Infrastructure link:
BS Æ MS3 or MS3 Æ BS
9 Ad hoc link:
MS1 Æ MS2
9
Concurrent Transmission Probability
Problem
‰ Where can MS1 set up an ad hoc link
without interfering the infrastructure
connection?
‰ What is the throughput performance of
the network consisting of both the
legacy infrastructure and CR-based ad
hoc links?
‰ CSMA/CA MAC protocol is adopted
9
a well-know and widely deployed MAC protocol in
unlicensed band.
Coexistence Region - Uplink
‰ The probability that both the SIRs in the
infrastructure and ad hoc links are larger than
the required SIR thresholds (zi, za)
MS1
Coexistence Region - Downlink
Coexistence Probability – UL case
Uplink Case
0.45
0.4
Insufficient
signal power in
infrastructure
link.
0.35
Existence Probability
0.3
0.25
0.2
0.15
Interference
from ad hoc
link.
SIRth=0dB(simulation)
SIRth=0dB(analysis)
SIRth=3dB(simulation)
SIRth=3dB(analysis)
0.1
0.05
MS1
0
0
10
20
30
40
50
60
Distance of MS3
70
80
90
100
‰ The coexistence probability in uplink case is up to 45%.
‰ The optimum point exists as the infrastructure user locates in R/2.
Throughput Performance
‰ The total throughput performance is 145% compared to the pure
infrastructure network.
Infrastructure Uplink Case
The more the allowable
ad hoc users, the less
the throughput
improvement due to the
more the collisions
among CR devices.
Normalized
Throughput
Total
throughput
1.8
1.75
1.7
1.65
1.6
1.55
1.5
100
0
20
40
# of users
50
60
80
Distance of MS3
100
0
Summary
‰ We dimension the region where a CR device can reuse the
frequency band to set up an ad hoc link with the primary
infrastructure link without the need of scanning the whole
spectrum.
¾
¾
The coexistence probability can be up to 45%.
It can achieve 145% throughput performance.
‰ With the help of location information from upper layer, an mobile
to mobile ad hoc CR user device could avoid scanning wide
spectrum and identify the spectrum opportunity.
Outline
Perspective on future research
‰ Background
‰ What is Cognitive Radio (spectrum overlay
‰ Cognitive MAC protocol Design
technique)?
‰ Challenges in Cognitive Radio
9 QoS Provisioning
9 Concurrent Transmission MAC Protocol
‰ Location Awareness for Dynamic Spectrum
‡ Game theory approach for spectrum
handoff for cognitive wireless system
‡ MIMO-based Cognitive Radio Networks
Access in CR networks
‰ Conclusions
Conclusion
Reference
1.
Li-Chun Wang and Anderson Chen, “On the Spatial
Coexistence of Infrastructure-Based and Ad Hoc
connections for a Cognitive Radio System,” submitted to
IEEE Trans. on Mobile Computing [Conference is available
at 1st International Conference on Cognitive Radio Oriented
Wireless Network and Communications (CROWNCOM), 2006,
Mykonos, Greece, June, 2006 ]
2.
Li-Chun Wang, Anderson Chen, and David W. L. Wei, “A
Cognitive MAC Protocol for QoS Provisioning in Ad Hoc
Networks,” submitted to IEEE Trans. on Vehicular
Technology. [Conference version is available at IEEE
Consumer Communications and Network Conference,
Cognitive Radio Workshop, Las Vega, Jan. 2007
3.
Li-Chun Wang, Chung-Wei Wang, Yin-Chih Lu, and ChuanMing Liu, “A Concurrent Transmission MAC Protocol for
Enhancing Throughout and Avoiding Spectrum Sensing
in Cognitive Radio,” IEEE Wireless Communications and
Networks Conference, Hong Kong, China, Mar. 2007
Cognitive Radio Is Hard!
Cognitive Network Is Harder!
Don’t miss your prime time!
Thank You!
lichun@cc.nctu.edu.tw
Current CR Research Activities
1. OFDM-based Spectrum Pooling
Techniques [Universität Karlsruhe
(TH), Germany, Weiss’04]
2. COgnitive Radio approach for usage of
Virtual Unlcensed Spectrum (CORVUS)
[UC Berkeley, Cabric & Brodersen’ 05]
3. IEEE 802.22 Wireless Reginal Area
Network (WRAN)
4. DIMSUMnet Project (Dynamic
Intelligent Management of Spectrum
for Ubiquitous Mobile network) [Bell
Lab, Buddhikot’05]
BACKUP
Current CR Research Activities
(Cont’d)
5. DRiVE Project (Dynamic Radio for IP
services in Vehicular Environments).
[Ericsson Eurolab Deutschland,
Germay , Xu&Toenjes’00
http://www.ist-drive.org/index2.html]
Follow-up project: OverDrive
(Spectrum Efficient Uni-and Multicast
Over Dynamic Radio Networks in
Vehicular Environments)
[http://www.comnets.rwthaachen.de/~o_drive/index.html]
Current CR Research Activities
(Cont’d)
6. Nautilus Project [UC Santa Barbara Zheng’05]
OTHER Cognitive Radio PLAYERS
‰ Vahid Tarokh – Information theoretical aspect
7. OCRA Project (OFDM-based Cognitive RAdio)
[Georgia Tech, Akyildiz and Li’06]
8. CWT (Cognitive Wireless Technology) [Virginia
Tech]
9. NCTU: NSC Project Access Technologies
and Resource Management for Cognitive
Radio Networks
¾
¾
Lead by Prof. C. J. Chang, W. H. Sheen, L. C. Wang
and C. Y. Huang
http://140.113.236.55/wise_lab/project/CR/cr.htm
‰
‰
‰
‰
of cognitive wireless networks
Harvard Univ
John Chapin -- Software-defined radios
Vanu, Cambridge, MA
Michael Honig -- Pricing algorithm for
spectrum sharing
Northwestern University
Joseph Mitola III -- Cognitive radios
Mitre, McLean, VA
Adam Wolisz -- Protocols for communications
networks
Technical University of Berlin, Germany