:
Jun-Hong Cui
Computer Science & Engineering
University of Connecticut

Many slides of this part are adapted from Debra Estrin, UCLA

 A sensor network is a network of integrated sensors embedded in the physical world
 Usually refer to wireless sensor networks
– Communication between sensors uses radio
 Three components of an integrated sensor
– Sensing
– Communication
– Computing
 Sensors are not dummy sensor anymore
 Smart sensors form autonomous net systems

• Many critical issues facing science, government, and the public call for high fidelity and real time observations of the physical world
• Networks of smart, wireless sensors can reveal the previously unobservable
• “The smarts” derives from coordination among the embedded devices to export information , not just data
• The technology will also transform the business enterprise, from the factory floor to the distribution channel

• Remote sensing transformed observations of large scale phenomena
• In situ sensing transforms observations of spatially variable processes in heterogeneous and obstructed environments
Embedded networked sensing will reveal previously unobservable phenomena
Red : Soil
Green : Vegetation
Blue : Snow
San Joaquin River Basin
Courtesy of Susan Ustin-Center for Spatial Technologies and Remote Sensing

• Embed numerous, low-cost, distributed devices to monitor and interact with physical world
• Deploy spatially and temporally dense, in situ, sensing and actuation
• Network these devices so that they can coordinate to perform higher-level identification and tasks
Requires robust distributed systems of thousands of devices.

Small, cheap, plentiful computing resources
SPEC (J. Hill):
4MHz/8bit, 3K/0K
Mica2Dot (Berkeley/Xbow):
8MHz/8bit, 4K/128K
Stargate (Intel/Xbow)
400Mhz/32bit, 64M/32M
Empty
Column
Filter
Filter &
Sensor
LC Column
Marine Algae Detector
(C Zhao) iMEMS Accelerometer
(Analog Devices)
Liquid Chromatograph
(YC Tai)
Small, cheap, plentiful sensing technologies



Physical environment is dynamic and unpredictable
Small wireless nodes have stringent energy, storage, communication constraints
WINS node
UCLA (1996)
Smart Dust
UCB (2000)


Large scale deployments call for processing and filtering of data close to sensor source
Embedded nodes must collaborate to report interesting spatio-temporal events
The network is the sensor!

Objectives Constraints
 Embeddable, low-cost sensor devices
 Robust, portable, self configuring systems
 Data integrity, system dependability
 Programmable, adaptive systems
 Multiscale data fusion, interactive access
Potentiometric Response for NO
3
-
Ion
320
280
Electrochemica l depo sition (constant current conditions) of polypyrro le dopped with nitrate onto carbon fibers substrate
240
200
160
120
80
1 2
Ca rbon fibers, 7  m dia meter each,
~ 20 -3 0 fibers, 1.2 cm depth
3 days after depositi on (Slope: 54 .3 mV, R
2
= 0 .99 99)
9 days after depositi on (Slope: 54 .4 mV, R
2
= 0 .99 99)
19 days a fter deposition (Slo pe: 5 2.6 mV , R
2
= 0.9 999)
5 6 3 4
-log(NO
3
-
)
 Energy
 Scale, dynamics
 Autonomous disconnected operation
 Sensing channel uncertainty
 Complexity of distributed systems

As the technology matures we will find wide-reaching applications in the built environment and throughout the business enterprise.

 The Earth is a water planet
– About 2/3 of the Earth covered by oceans
• Uninhabited, largely unexplored
• A huge amount of (natural) resources to discover
 Many potential applications
– Long-term aquatic monitoring
• Oceanography, marine biology, deep-sea archaeology, seismic predictions, pollution detection, oil/gas field monitoring …
– Short-term aquatic exploration
• Underwater natural resource discovery, hurricane disaster recovery, anti-submarine mission, loss treasure discovery …

 Desired properties
– Unmanned underwater exploration
– Localized and precise data acquisition for better knowledge
– Tetherless underwater networking for motion agility/flexibility
– Scalable to 100 ’ s, 1000 ’ s of nodes for bigger spatial coverage

Submarine Detection
Data Report
Acoustic
Radio
Buoys
Sonar Transmitter

 Existing Approaches
– Active or passive sonar
– Cons: submarine anti-detection techniques (e.g., sonar absorption) make them less-effective
 Using UWSN
– Collaborative detection
• Multiple sensors, and/or multi-modal data
– Large coverage
– Timely reporting
– High reusability

Estuary Monitoring
Fresh
Buoyancy
Control
Fresh Water Current
Salty Water Current
Buoyancy
Control
Salty

 Existing Approaches
– Ship tethered with chains of sensors moves from one end to the other
– Cons: no 4D data, either f(x, y, z, fixed t), or f(fixed (x, y, z), t); and cost is high
 Using UWSN
– Easily get 4D data, f(x, y, z, t), sensors move
– Reduce cost significantly
– Increase coverage
– Have high reusability

 Sensor node system design
– Sensing, computing, communication integration
– Power management: energy saving, life time
 Autonomous network system design
– Communication, multiple access
– Routing, forwarding, reliable transfer
– Localization, synchronization
– Security, robustness
– Energy efficiency

 Applications and data management
– Application classification & characterization
– Data sampling, structure, storage
 Collaborative estimation & detection
– Data fusion, dissemination, tracking
 Modeling, simulation, evaluation
– Network simulator
– Sensor node simulator
 Hardware, middleware, software design

Environmental constraints
Application requirements
Energy consumption model
Sensor node design
Network design
Resource management
Other design components
Lifetime estimation model
UWSN system parameters

 Narrow bandwidth channels
– High-frequency waves rapidly absorbed by water
 radio not applicable in water
– Must use acoustic channels - low bandwidth, fading
 High attenuation
– Bandwidth X Range product = 40 Kbps x Km
– Very low compared to RF channels (1:100)
• 802.11b/a/g yields up to 5Mbps x Km
 Very slow acoustic signal propagation
– 1.5x10
3 m / sec vs. 3x10 8 m / sec
– Causes large propagation delay

Courtesy: Kilfoyle & Baggeroer
Reported by
Kaya&Yauchi,Oceans'89
Jones et al.,Oceans'97
Capellano et al.,Oceans'97
Modulation Method Bandwidth Bandwidth Carrier Data Rate Range
16QAM
DPSK
BPSK
125kHz
10kHz
0.2kHz
1000kHz
50kHz
7kHz
500kbps 60m
20kbps 1km
0.2kbps
50km

 UnderWater Acoustic (UW-A) channel:
– Narrow band: hundreds of kHZ at most
– Huge propagation latency
– High channel error rate
 Random topology and sensor node mobility
(1--1.5m/s due to water current)
– Existing protocols in terrestrial sensor networks assume stationary sensor nodes;
– In mobile sensor networks, these protocols weakened
 Mobility & UW-A channel limitations open the door to very challenging networking issues

 UWSNs must require:
– Reliable data transfer (tolerating high error-prone acoustic channels)
– Efficient data delivery (should be energy-efficient)
– Localization (for geo-routing or meaningful data)
– Time synchronization (for sleep cycle schedule, multiple access protocol schedule, etc)
– Efficient multiple access (sensors are densely deployed)
– Efficient acoustic communication (improving data rate)
 Design Objective:
– Build efficient, reliable, and scalable UWSNs

 High-precision localization is a must for 4D sampling
 Current approach: UAV interrogate fixed references (0.5m)
 Architecture for estuary monitoring: underwater GPS
Surface buoys collaborative localization via radio links sensors self-localization via acoustic links
Optional ancored reference point
26

 Localize large number of nodes for routing protocols
 Propose a hierarchical localization approach
Surface buoy Anchor nodes sensor nodes
 Anchor Node
Localization
 Underwater GPS
 Ordinary Node
Localization
 3-D Euclidean
Distance Estimation
 Recursive Location
Estimation
Mobility prediction is key in mobile UWSNs

 UWSN is challenging and promising new area
– Requires interdisciplinary efforts from
• Environmental engineering
• Acoustic communication
• Signal processing
• Network design
 Future Work
– A long to-do list …
– Your active participation is warmly invited
• Application characterization, environmental modeling, water tracking, localization, sensing …

 Sensor Network and Systems research
– Jun-Hong Cui, Computer Science & Engineering (Director)
– Yunsi Fei, Electrical & Computer Engineering
– Jerry Zhijie Shi, Computer Science & Engineering
– Bing Wang, Computer Science & Engineering
– Peter Willett, Electrical & Computer Engineering
– Shengli Zhou, Electrical & Computer Engineering (Co-director)
 Algorithmic and Performance support
– Reda Ammar, Computer Science & Engineering
– Lanbo Liu, Civil & Environmental Engineering
– Sanguthevar Rajasekaran, Computer Science & Engineering
 Context and Applications consultation
– Amvrossios Bagtzoglou, Civil & Environmental Engineering
– Thomas Torgersen, Marine Sciences

 Equipment List:
– Acoustic modem
– Underwater speaker
– Hydrophone
– Sound mixer
– Sound receiver
– Speaker/microphone
– Aquarium

 Designed and Implemented by WHOI (Woods Hole Oceanographic
Institution)
 A Low-power
Acoustic Modem
 Based on the
TMS320C5416
DSP from TI

 Control-1 150 Watt
Two-Way Loudspeaker
System
– Good performs in recording studios
– Low distortion reproduction
– Frequency Range: 70 Hz - 20 kHz
 Sony STRDE197
Stereo Receiver
 Sennheiser MKE 300
Microphone

 Frequency range: 200 Hz to 32 KHz
 Directional at higher frequencies
 A completely passive, non-powered device
 Can be used as an air speaker or a receive hydrophone

 Output:
– 300mW, short-circuit-proof
– 3.5mm (mini) phone jack
 Power Requirements:
– 7mA quiescent current
 Usable Frequency
Response:
– 20Hz - 100KHz
 Polar Response:
– Omni directional

 Ultra-Pure Sound and Crystal-Clear Audio
 99 special sound effects:
– Reverbs
– Delays
– Tube distortion
– And More!
 24 channels
 Could simulate different underwater environments

 Distance between the underwater speaker and hydrophone: 1 meter