Transcript Wireless Sensor Networks - Systems and Computer Engineering

Wireless Sensor Networks
RCTI Seminar Day Presentations
Roshdy Hafez
Thomas Kunz
Marc St.-Hilaire
Ionnis Lambadaris
Richard Yu
Roshdy Hafez
Systems and Computer Engineering
Thomas Kunz
Professor and Director
Technology Innovation Management Program
Mobile Computing Group
Facilitate the development of innovative next-generation mobile
applications on resource-constraint, mobile devices
Develop the required network architectures (MANETs, wireless
mesh networks, wireless sensor networks)
Research into network protocols (MAC, routing, Mobile IP, QoS
support, transport), and middleware runtime support
Licensed technology to EION Inc. in 2005 (Adaptive Intelligent
Router)
Research funded by federal (NSERC) and provincial granting
agencies (OCE, NCIT), as well as industry
– Worked with Bell, Nortel, Motorola in the past
– Currently cooperating with CRC, Alcatel-Lucent
High-Level Architecture: multiple WSN, fixed Core
(Examples: surveying multiple airports, border crossings, etc.)
Wireless Sensor Networks:
dynamic retasking, new
sensor types/data, improved
algorithms and protocols
Fixed Networking:
distribute sensor data to
(different) recipients, discover
sensors and their capabilities
Event
collection &
presentation
Monitoring data processing
Event dissemination
1st responder notification
XML Routed Network
sensor data collection
and archive:
information made
available via web
services
IP
IP Router
Base Station
XML
Router
Monitored Area
Sensor
Core Functionality: Clock Synchronization, Localization
Clock sync is critical at many layers
– Beam-forming, localization, distributed DSP, MAC
– Tracking; data aggregation & caching
Similarly, localization is fundamental
– Routing, security
– Tracking; data aggregation & caching
t=2
t=1
t=3
t=0
Localization
Key requirements: high accuracy, no additional hardware (GPS, etc.),
support fast deployment (minimum # of anchors), range-free or range-based
Another important point: should work well for typical “mission-critical” deployments
10
10
10
8
8
8
6
6
6
4
4
2
0
2
0
4
0
2
4
6
8
(a)
Random topology, 200 nodes
10
2
0
5
(b)
C-shaped network,
160 nodes
10
0
0
5
(c)
10
Uniform grid (with
small placement errors)
100 nodes
Localization: Cooperative Localization,
based on Curvilinear Component Analysis
(a) Range free: 3 anchor nodes
(a) Range-based: 3 anchor nodes
0.5
0.15
0.1
0.05
0
10
15
20
25
Connectivity
(b) Range based: 10 anchor nodes
0.3
0.2
30
0.1
0.05
0
5
10
15
20
Connectivity
25
5
10
15
20
25
Connectivity
(b) Range free: 10 anchor nodes
30
0.4
MDS-MAP(P,R)
CCA-MAP
0.15
MDS-MAP(P,R)
CCA-MAP
0.4
0.1
5
0.2
Median Error (r)
Median Error (r)
MDS-MAP(P,R)
CCA-MAP
Median Error (r)
Median Error (r)
0.2
30
MDS-MAP(P,R)
CCA-MAP
0.3
0.2
0.1
0
5
10
Results for Random Network Topology
15
20
Connectivity
25
30
Clock Synchronization
mutual, low overhead, compatible withWiFi, WiMax, Zigbee standards (i.e., based
on periodic beacons)
key idea: adjust slope of local clocks, rather than timestamp value -> converge over time
c.d.f of max time difference in a 5x5 network
using the IEEE 802.11 TSF
Max time difference in a 5x5 network using
CSMNS
Steps Forward
 Defined and evaluated fundamental algorithms through simulations
 Plan to implement and evaluate them in a real testbed
 Additional research questions
– Localization:
Optimal anchor locations (non-trivial and non-obvious)
Apply NN structure to track mobile sensors
Reduce computational complexity
Bound worst-case performance
– Synchronization:
Use external clock references
Reflect hierarchical network structure
 Ongoing: work on fixed-network aspects, gateway to interconnect WSN
and core, XML-based description and discovery, etc.
Marc St-Hilaire
School of Information Technology
Wireless Sensor Networks (WSN)
Research in planning algorithms (both static & dynamic)
– How to design new WSN in a cost effective way
– How to update an existing WSN infrastructure
– Organisation (re-organisation) of the nodes to maximize the life time of the network
Research on network protocols
– Routing scheme with different objectives
Save energy, minimise delay or combination
Re-organise the route in case of node/link failure
– Correlation of events both in space and time
Clock synchronisation
Localization algorithm
Wireless Sensor Networks (WSN)
Research on data association
– How to follow multiple moving targets such as in military applications, border
defence and so on.
Research on data aggregation/fusion
– Aggregate data in order to save bandwidth, computing resources, battery life, etc.
Ioannis Lambadaris
Systems and Computer Engineering
Overview: Research/Academic Interests
John Lambadaris
Associate Professor
Department of Systems and Computer
Engineering
Carleton University
Ottawa, Ontario K1S 5B6
email: [email protected]
tel: (613) 520-2600 x1974
Performance Analysis of Computer Communication Networks

Congestion control of IP networks, Differentiated services and Quality of Service

Resillient Packet Ring protocols and performance evaluation

Resource allocation and Quality of Service in optical networks

Real time packet content inspection engines
Security

Endpoint-Driven Intrusion Detection and Containment of Fast Spreading Worms in
Enterprise Networks
Mobile/Wireless Networks

High Speed Downlink Packet Access (HSDPA)
Sensor and Ad-Hoc Networks

Zigbee/IEEE 802.15.4 networking
Practical Design for wireless sensor nodes

Design, performance analysis and prototyping of nodes based on popular wireless
transceivers such as TI/Chipcon (CC1100, CC1110), Freescale semiconductors
(MC13201-2-3 ), Cypress Semiconductors (CYRF69103, CYRF69213)
Distinctions:

Ontario Premiers Excellence Award 1999 -- Carleton Research Achievement Award
2000-01.
Patents: 20060089113 - Radio control receiver system for multiple bands, frequencies and modulation protocol coverage.
Authors: John Lambadaris, A. Elahi and J. Perez
Topics to address:
•High Speed Downlink Packet Access (HSDPA) systems
•Sensor/wireless ad-hoc networks
-Node Location Estimation
-Low Bit rate video for surveillance
Optimal Scheduling in High Speed Downlink Packet Access
(HSDPA)
•Objective
-To find the optimal
scheduling policy that controls the allocation of the timecode resources.
•An optimal policy should be:
-Fair; Divide the resources fairly between all the active
users.
-Maximize the overall cell throughput.
-Provide channel aware (diversity gain) and high speed
resource allocation.
Optimal Scheduling in HSDPA: Analysis and
Validation
• Methodology
-Markov Decision Processes and Dynamic Programming (two user analysis)
-OPNET based simulations for verification
Optimal policy (two user case)
Comparison with heuristic policies
Optimal Scheduling in HSDPA: Further research
-Realistic channel modeling
-Packet retransmissions
-Scalability issues
-Extension to more than two users
Recent publications:
Hussein Al-Zubaidy, Ioannis lambadaris, Code Allocation Policy Optimization in HSDPA Networks
Using FSMC Channel Model, IEEE Wireless and Networking Conference (IEEE WCNC), March 31April3, 2008.
Sensor Location Estimation: Problem Statement
• The sensor localization problem.
– Given a set of sensors deployed in a field, in which
some of them are anchors and the remaining are
unknown sensors, we may want to estimate the nodes
positions of the unknown sensors.
• Anchors: Nodes that know their positions.
• Unknown sensors: Nodes that do not know their
positions.
Sensor Location Estimation:
Range-based and Range-free algorithms
•In order to study the sensor localization problem, researchers
have proposed schemes that lie on one of the following
categories:
–Range-based algorithms rely on computing point-topoint distance estimates.
–Range-free algorithms propose solutions without the
availability of inter-distance measurements.
•Our hybrid approach: We will use a range-free approach
coupled with a range-based refinement.
Sensor Location Estimation:
APIT Algorithm
a is an unknown sensor.
A,B,C,D are audible anchors for a.
Step:
1. Generation of triangles.
3 combinations from the set of
4 audible anchors = 4 triangles
-> {ABC,BCD,ACD,ABD}
2. Acquisition of beacon
information.
3. APIT Test
4. APIT Aggregation.
5. Position estimation (COG).
Sensor Location Estimation: Simulation Setup
Random distribution
Sparse Networks
Random distribution
Black nodes ->anchors,
White nodes -> unknown sensors
Deterministic distribution of anchors
Dense
networks
A Propagation Model for Sensors: RIM
(Radio Interference Model)
DOI (Degree of Irregularity)
parameter
Maximum path loss percentage
variation per unit degree change
in the direction of radio
propagation.
RIM Model
Model that introduces the
DOI parameter.
Anisotropic model.
Radio variations depend with
both distance and direction.
Sensor Location Estimation: Results
• N=40, R=1.5 [m]
• M=200, N=40, R=1.5 [m]
DOI=0.1
DOI=0.7
Sensor Location Estimation: Further research
•Time varying interference patterns
•Extensions of the location algorithms to include obstacles
(e.g. terrain irregularities) between nodes
•Complexity and scalability of the algorithms
•Extensions to include node/sensor mobility
Low bit-rate Video Transmission over Wireless Zigbee Networks
Challenges
• Video application requirements




High data rate for high quality (compression is used)
Bandwidth-efficient codecs are the most computationally intensive
Limitations of Zigbee networks



Low Power (Battery operated)
Maximum nominal rate for IEEE 802.15.4 standard is 250 kbps
Realistic throughput is much lower (CSMA/CA, overhead, multi-hop, etc.)
Video applications may be implemented over Zigbee



Using advanced video encoders, video segmentation and rate-control algorithms
Using the multiple channels available in the IEEE802.15.4 and using multiple NICs
Using MDC and multi-hopping over multi-channel multi-interface network topologies
Recent Publication: Ahmed Zainaldin, Ioannis Lambadaris, Bis Nandy Adaptive Rate Control MPEG4 Video Transmission over
Wireless Zigbee Networks, IEEE International Conference on Communications (ICC), May 19-23 2008
Solutions for Video Transmission over Zigbee Networks
1.
Rate Control Variable bit-rate over Wireless Zigbee Networks (RCVBR)
Video
Source
MPEG-4
Encoder
b
R(n)
r
Packetizer
Q
Rate
Controller
2.
Region of Interest (ROI) Encoding
3.
Multi-channel Multi-radio over Wireless Zigbee Networks
1
1
11
m
4.
11
m
Multiple Description Coding (MDC) over a multi-channel multiinterface Zigbee networks
Summary: Research expertise and personnel
•Simulations, traffic modeling and performance analysis
-NS-2 and OPNET based simulations
•Matlab computations for propagation and interference models
•Prototyping sensor node/development from concept to manufacturing (PCB design,
firmware programming, RF design)
•Personnel: Faculty, graduate students, research associates and a group of
professional contractors
Secure Wireless Biosensors Networking for
Authentication and Life Support of Field
Personnel
Richard Yu
RCTI, Carleton University
Helen Tang and Peter Mason
DRDC - Ottawa
 Military tactical mobile ad hoc networks (MANETs) challenge
security design.
 As the front line of defence, authentication is the core
requirements for integrity, confidentiality and non-repudiation
in networked centric warfare.
 Biometrics from biosensors provide some promising solutions to
the authentication problems.
Fingerprint
Finger vein
Iris
Face
Cardio-based
Retina
Voice
 Patient/citizen centered healthcare based on wireless biosensors
•
A unified framework approach
Sensor data
Multimodal
Biometrics
User
authentication
Encryption
Physiological status
monitoring
 Research: Wireless networking for biosensors, biometric-based
authentication for tactical MANET, biosensor data processing,
biosensor scheduling and management.