Sensor Network Research: Emerging Challenges - PowerPoint PPT Presentation

Loading...

PPT – Sensor Network Research: Emerging Challenges PowerPoint presentation | free to download - id: b0e57-ZTU2Y



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Sensor Network Research: Emerging Challenges

Description:

169 motes, 13x13 grid, 2 ft spacing, open area, RFM radio, simple CSMA. Can't just ... Unlike Internet, the location of nodes in time and space is essential for local ... – PowerPoint PPT presentation

Number of Views:78
Avg rating:3.0/5.0
Slides: 52
Provided by: debo129
Learn more at: http://www.cens.ucla.edu
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Sensor Network Research: Emerging Challenges


1
Sensor Network Research Emerging Challenges
  • Deborah Estrin
  • Director, Center for Embedded Networked Sensing
  • http//cens.ucla.edu/
  • Professor, UCLA Computer Science Department
  • destrin_at_cs.ucla.edu
  • http//lecs.cs.ucla.edu/estrin Acknowledge
  • NSF, DARPA, Intel, Sun
  • Student and Faculty collaborators at UCLA,
    USC-ISI, UCB, Intel

2
Embedded Networked Sensing (ENS) A Transforming
Technology
  • Imagine if
  • High-rise buildings in Los Angeles were able to
    detect their own structural faults (e.g., weld
    cracks or plumbing infrastructure)
  • Belmont school could reliably measure toxic
    levels at very low concentrations, and trace
    contaminant transport back to its source
  • Buoys along the coast could alert surfers,
    swimmers, and fisherman to dangerous bacterial
    levels
  • An earthquake-rubbled building could be
    infiltrated with robots and sensors to locate
    signs of life and evaluate structural damage
  • We could infuse complex and endangered ecosystems
    with a plethora of chemical, physical, acoustic,
    and image sensors to track global change
    parameters continuously.
  • Dangerous bacterial and contaminant levels could
    be detected on the farm through dense sampling,
    instead of in the market through sparse sampling

3
Embedded Networked Sensing Potential
  • Micro-sensors, on-board processing, wireless
    interfaces feasible at very small scale--can
    monitor phenomena up close
  • Enables spatially and temporally dense
    environmental monitoring
  • Embedded Networked Sensing will reveal
    previously unobservable phenomena

Seismic Structure response
Contaminant Transport
Ecosystems, Biocomplexity
Marine Microorganisms
4
Enabling Technologies
Embed numerous distributed devices to monitor and
interact with physical world
Network devices to coordinate and perform
higher-level tasks
Embedded
Networked
Exploit collaborative Sensing, action
Control system w/ Small form factor Untethered
nodes
CENS
Sensing
Tightly coupled to physical world
Exploit spatially and temporally dense, in situ,
sensing and actuation
5
Seismic Structure Response The Problem
  • Earthquake impact Injury, loss of life,
    financial loss.
  • Interaction between ground motions and
    structure/foundation response not well
    understood.
  • Current seismic networks not spatially dense
    enough
  • to monitor structure deformation in response to
    ground motion.
  • to sample wavefield without spatial aliasing.
  • ENS will provide field data at sufficient
    densities to develop predictive models of
    structure, foundation, soil response.

6
Seismic Structure Response Field Experiment
  • 38 seismometers in 17-story steel-frame Factor
    Building.
  • 100 free-field seismometers in UCLA campus
    ground at 100-m spacing

??¾¾¾¾¾¾ 1 km ¾¾¾¾¾¾?
7
Contaminant Transport The Problem
  • Pollutants place humans and ecosystems at risk.
  • Sparse, labor-intensive observations, linked by
    un-validated transport models, result in shaky
    exposure assessments…propagate to even shakier
    risk assessments
  • ENS will render exposure assessment
    trivial, and pave the way to meaningful risk
    assessment.

8
ENS Research for Contaminant Transport
  • Environmental Micro-Sensors
  • Sensors capable of recognizing phases in
    air/water/soil mixtures.
  • Sensors that withstand physically and chemically
    harsh conditions.
  • Microsensors.
  • Signal Processing
  • Nodes capable of real-time analysis of signals.
  • Collaborative signal processing to expend energy
    only where there is risk.

9
Ecosystem Monitoring The Problem
  • Biodiversity and the supporting ecosystems
  • Contribute trillions of to global economy
  • Changing rapidly with catastrophic consequences
  • Biological and environmental complexity of
    ecosystems not well understood (climate change,
    exotic species introduction, …)
  • Mitigation efforts costly and difficult to assess
  • Environmental laws, Land acquisition
  • ENS systems will transform the study of
    biocomplexity and global change by making it
    feasible to record detailed combinations and
    complex patterns of organism interaction within
    dynamic ecosystems.

10
(No Transcript)
11
ENS for Ecosystems Monitoring
  • Dense network of physical, chemical sensors in
    soil and canopy
  • Multimedia sensors in natural habitats and
    artificial cavities (nest boxes).
  • Wireless due to climate, terrain, thick
    vegetation.
  • Adaptive Self-Organization for reliable,
    long-lived, operation in dynamic,
    resource-limited, harsh environment.
  • Mobility for deploying scarce resources (e.g.,
    high resolution sensors).

Virtual field observations
12
ENS enabled by Networked Sensor Node Developments
LWIM III UCLA, 1996 Geophone, RFM radio, PIC,
star network
AWAIRS I UCLA/RSC 1998 Geophone, DS/SS Radio,
strongARM, Multi-hop networks
Sensor Mote UCB, 2000 RFM radio, PIC
Medusa, MK-2 UCLA NESL 2002
  • Predecessors in
  • DARPA Packet Radio program
  • USC-ISI Distributed Sensor Network Project (DSN)

13
ENS New Design Themes
  • Long-lived systems that can be untethered
    (wireless) and unattended
  • Communication is the primary consumer of scarce
    energy resources
  • Every bit transmitted brings a sensor node
    one moment closer to death (G. Pottie)
  • Leverage data processing inside the network
  • Exploit computation near data to reduce
    communication
  • Collaborative signal processing
  • Achieve desired global behavior with localized
    algorithms (distributed control)
  • Self configuring systems that adapt to
    unpredictable environment
  • Dynamic, messy (hard to model), environments
    preclude pre-configured behavior

14

The network is the sensor
  • Requires robust distributed systems of
  • thousands of
  • physically-embedded,
  • unattended,
  • and often untethered,
  • devices.
  • (MangesSmith, Oakridge Natl Labs, 10/98)

15
ENS Architecture Drivers
Drivers
Research Areas


Adaptive Self-Configuring Systems
Varied and variable environments


Distributed Signal and Information Processing
Energy and scalability


Sensor Coordinated Actuation
Heterogeneity of devices

Embeddable Microsensors
Smaller component size and cost
16
Long-Lived, Self-Configuring Systems
  • Irregular configurations
  • Network topology changes over time
  • Hand configuration will fail -- scale, and
    variability
  • Solution local adaptation and redundancy
  • Challenges
  • Localization
  • Time Synchronization
  • Calibration
  • Information aggregation and storage
  • Event detection
  • Programming model!

Local sensors
17
Context for creating a topology connectivity
measurement study (Ganesan et al)
Cant just determine Connectivity clusters
thru geographic Coordinates… For the same
reason you cant determine coordinates
w/connectivity
Packet reception over distance has a heavy tail.
Non-zero probability of receiving packets at
distances much greater than the average cell range
169 motes, 13x13 grid, 2 ft spacing, open area,
RFM radio, simple CSMA
18
Fine Grained Time and Location
  • Unlike Internet, the location of nodes in time
    and space is essential for local and
    collaborative detection
  • Fine-grained localization and time
    synchronization needed to detect events in three
    space and compare detections across nodes
  • GPS provides solution where available (with
    differential GPS providing finer granularity)
  • Acoustic or Ultrasound ranging and
    multi-lateration algorithms promising for non-GPS
    contexts (indoors, under foliage…)
  • Fine grained time synchronization needed to
    support ranging

19
Tiered System Design IPAQs and
UCB/Intel/Crossbow Motes
  • Localization using acoustic ranging
  • IPAQs listen for acoustic chirps from motes
    (buffer time series - mote cant do this)
  • Run matched filter, record time diff btwn emit-
    and receive-time of coded sequence
  • Share ranges with each other trilaterate
  • IPAQs range to each other to self configure
  • Time sync based on multi-hop Reference Broadcast
    Synchronization
  • Allows computation of acoustic time-of-flight
  • One IPAQ has a MoteNIC to sync mote and IPAQ
    domains

20
Time Sync Enables Beamforming
Localization Experiment
H. Wang et. al.
21
Collaborative Signal Processing and Active
Databases
robot
Coherent combining
human observer
transmit decision
cooperative processing
high resolution processing, main database,
policy adaptation
query for more information
queue data for internal processing, queries from
higher levels
22
Programming Challenge
  • How do we task a 1000 node dynamic sensor
    network to conduct complex, long-lived tasks ??
  • Identify Spatio-temporal, multi-modal, events
  • Scalability
  • Energy constrained…Communication constrained

23
System Architecture Current state of the art
and community consensus…
  • Its a Database!…
  • NO, its a wireless Ad Hoc Network!…
  • NO, its an Internet!…
  • NO, its a Neural Net!…
  • NO, its an Parallel computer!…
  • NO, its an Distributed system!…

24
Why cant we simply adapt Internet protocols and
the end to end architecture?
  • Internet routes data using IP Addresses in
    Packets and Lookup tables in routers
  • Humans get data by naming data to a search
    engine
  • Many levels of indirection between name and IP
    address
  • Works well for the Internet, and for support of
    Person-to-Person communication
  • Embedded, energy-constrained (un-tethered,
    small-form-factor), unattended systems cant
    tolerate communication overhead
  • Name the data, not the nodes even at the lowest
    levels of the system.
  • ENS systems raise many new technical challenges

25
Its NOT just an Internet Directed Diffusion
Data Centric Routing
  • Basic idea
  • name data (not nodes) with externally relevant
    attributes
  • Data type, time, location of node, SNR, etc
  • diffuse requests and responses across network
    using application driven routing (e.g., geo
    sensitive or not)
  • optimize path with gradient-based feedback
  • support in-network aggregation and processing
  • Data sources publish data, Data clients subscribe
    to data
  • However, all nodes may play both roles
  • A node that aggregates/combines/processes
    incoming sensor node data becomes a source of new
    data
  • A sensor node that only publishes when a
    combination of conditions arise, is a client for
    the triggering event data
  • True peer to peer system
  • Implementation defines namespace and simple
    matching rules with filters
  • Linux (32 bit proc) and TinyOS (8 bit proc)
    implementations

26
Diffusion as a construct for in-network
processing
  • Nodes pull, push, and store named data (using
    tuple space) to create efficient processing
    points in the network
  • e.g. duplicate suppression, aggregation,
    correlation
  • Nested queries reduce overhead relative to edge
    processing
  • Complex queries support collaborative signal
    processing
  • propagate function describing desired
    locations/nodes/data (e.g. ellipse for tracking
    (Zhao et al))
  • Interesting analogs to emerging peer-to-peer
    architectures
  • Build on a data-centric architecture for queries
    and storage

27
A more general look at Data Centric vs. Address
Centric approach (Krishnamachari et al.)
  • Address Centric
  • Distinct paths from each source to sink.
  • Data Centric
  • Support aggregation in the network where
    paths/trees overlap
  • Essential difference from traditional IP
    networking
  • Building efficient trees for Data centric model
  • Aggregation tree On a general graph if k nodes
    are sources and one is a sink, the aggregation
    tree that minimizes the number of transmissions
    is the minimum Steiner tree. NP-complete….Approxim
    ations
  • Center at Nearest Source (CNSDC) All sources
    send through source nearest to the sink.
  • Shortest Path Tree (SPTDC) Merge paths.
  • Greedy Incremental Tree (GITDC) Start with path
    from sink to nearest source. Successively add
    next nearest source to the existing tree.

28
Comparison of energy costs
Data centric has many fewer transmissions than
does Address Centric independent on the tree
building algorithm.
Address Centric Shortest path data centric Greedy
tree data centric Nearest source data
centric Lower Bound
29
ENS systems can be viewed as Databases
  • Sensor networks are capable of producing massive
    amounts of data
  • Sensor networks should be able to
  • Accept queries for data
  • Respond with results
  • Users will need
  • An abstraction that guarantees reliable results
  • Largely autonomous, long lived network
  • Efficient organization of nodes and data will
    extend network lifetime
  • Database techniques already exist for efficient
    data storage and access

30
What Wed Like to Do Sensor Database System
  • Sensor Database System that supports distributed
    query processing over a sensor network, instead
    of traditional warehousing approach

Sensor DB
Sensor DB
Sensor DB
Sensor DB
Sensor DB
Front-end
Sensor DB
Sensor DB
Sensor DB
Sensor Nodes
31
State of the Art The Tiny Aggregation (TAG)
Approach (Taken directly from S. Madden, UCB)
  • Push declarative queries into network
  • Impose a hierarchical routing tree onto the
    network
  • Divide time into epochs
  • Every epoch, sensors evaluate query over local
    sensor data and data from children
  • Aggregate local and child data
  • Each node transmits just once per epoch
  • Pipelined approach increases throughput
  • Depending on aggregate function, various
    optimizations can be applied

32
Distributed Representation and Storage
  • Interpretation of Spatially Distributed Data
  • Interpret data at different spatial and temporal
    extents. Per-node processing alone is not enough
  • Index data for easy temporal and spatial
    searching
  • Pattern-Triggered Data Collection
  • Multi-resolution data storage and retrieval
  • Data Centric Protocols and In network Processing
    goal
  • Network does in-network processing based on
    distribution of data
  • When data changes, portion of network updates
    understanding of data--histograms, wavelets…
  • Queries automatically directed towards nodes that
    maintain relevant/matching summarized data

33
Building Block How should nodes summarize
data? (Bien, UCLA)
  • World is one big data distribution that changes
    over time
  • Sensor nets monitor changes in environmental data
  • Communication should be driven solely by these
    changes
  • Histograms are a useful way to summarize data and
    track changes
  • Queries for statistics
  • Show degree of data variation in an area
  • Queries for maps
  • Geographically show data variation
  • Provide approximation of map

Y Location
X Location
34
Sampling Spatially-Correlated Data
  • Sampling data in uniformly random manner not
    appropriate
  • Higher resolution sampling useful where more is
    going on (i.e. more variance in data values)
  • Noise in data might yield inaccurate high-level
    view of data
  • Measuring Dynamism of Data with Statistical
    Pre-processing
  • Each node tracks variance of its own data
  • Nodes locally exchange histograms representing
    sensor values in local area
  • Each node compares amounts of error in histogram
    before and after adding its own sensor
    value--contributes if information added
  • When nodes sensor value changes significantly
    enough, recalculates its datas variance

35
Multi-Resolution Storage Architectures
DIMENSIONS
  • Per-Node Temporal Data Processing
  • Progressive encoding more lossy storage of
    older data. (Culler)
  • Design Principle guarantee lossless
    multi-resolution data collection within time T of
    event
  • Spatial Data Processing
  • Nodes form logical spatial hiearchy.
  • At each level in hierarchy, further summarization
    of data

increasing spatial resolution
increasing temporal resolution
Time
Ganesan et al.
36
Building Block Data Centric Storage (Ratnasamy,
ICSI)
  • Some queries (e.g. local maxima) not spatially
    correlated.
  • Require storage scheme that indexes on data
    semantics as well as location.
  • Data Centric Storage hashes on event name to
    support event (info) discovery (rendezvous)

37
An Example Semantic Hierarchy using
DCS (Greenstein, UCLA)
  • The wider the geographical extent a node indexes,
    the more data there is to index.
  • Given limited storage, tighter semantic
    constraints used for nodes that index wider areas.

38
Contour Gradient Finding Background
  • Contours and gradients help identify regions of
    interest in a data set.
  • Finding boundary of phenomenon important to
    understanding what it could effect.
  • Applications
  • Weather prediction
  • Contamination transport
  • Topography

Isothermic picture of El Nino (NASA)
39
Edge Detection
  • Extensively studied in Computer Vision
  • Canny, Binford, Nalwa
  • How is this different ?
  • Distributed context
  • Irregular Pixel placement.
  • Communication is not free.

40
Distributed Edge Detection Algorithm example
  • Nodes exchange data with their neighbors.
  • Each node calculates its gradient.
  • A node having a large gradient asks neighbors
    if they have also observed large gardients.
  • Nodes having sufficiently large gradients pass
    the message along.
  • Nodes having small gradients send back messages
    indicating the extent.
  • Once the extent is known, a node decides whether
    the edge fits the criteria.

Edge detection continues outward from there.
41
First Things First…Calibration (Bychkovskiy-Megeri
an, UCLA) and (Whitehouse, UCB)
  • Storage, forwarding, aggregation, triggering
    useless unless data means something
  • Sensors must be calibrated
  • identical stimuli should produce identical
    responses
  • Traditional solution
  • In-factory calibration
  • Why in-factory calibration is not sufficient?
  • calibration must account for coupling of sensors
    to environment

Factory Calibrated Sensors T0
72º
72º
72º
72º
72º
72º
Factory Calibrated Sensors Later
62º
70º
72º
71º
72º
72º
Dust
42
General Calibration Techniques
  • In factory (pre-deployment)
  • Pros low cost(), easy to perform
  • Cons no correction for physical coupling
  • Active in-situ Mobile actuator elements
  • Pros corrects for coupling
  • Cons limited applicability, not scalable,
    expensive
  • Passive in-situ Distributed Self Calibration
  • Pros corrects for coupling, scalable, low
    cost()
  • Cons requires higher sensor density, consumes
    energy, cant cover all cases

43
Distributed Self-Calibration (Bychkovskiy-Megeria
n, UCLA)
Additional sensors
  • Operational Requirements
  • understand application requirements
  • e.g., need 1 sensor/m2 and 0.1º accuracy
  • Over-deployment
  • Calibration Techniques
  • derive local mappings in situ
  • use tiered sensing
  • compensated sensors as a distributed baseline
  • Detect local and systematic bias (dust,
    obstructions, …)

t0
t1
44
CENS Systems under construction http//cens.ucla.e
du
  • Biology/Biocomplexity
  • Microclimate monitoring (for James Reserve using
    UCB Weather board for MICA (Mainwaring, Intel))
  • Triggered image capture (for James Reserve
    Mosscam and Middleschool Biology observatory)
  • Canopy-net (for Wind River Canopy Crane Site)
  • Environmental monitoring component of Smithsonian
    Tropical Research Institute (STRI) tracking
    facility
  • Geophysics
  • 100-node field network, 100 meter spacing
  • Correlated with structure response in
    USGS-instrumented Factor Building w/ augmented
    wireless sensors
  • Laboratory scale Contaminant Transport and Marine
    Microorganisms Observations
  • Laboratory instrumentation for 100 point
    observations in tank
  • Sensor coordinated actuators for sample
    collection

45
Sensor-Coordinated Actuation
  • Robotic elements extend system lifetime and
    capabilities
  • Fill in topology gaps
  • Recharge
  • Sample collection (physical, image…)
  • Static elements increase speed, efficiency,
    practicality of robotic elements
  • Navigation through data/sample space
  • Extend effective view of robot to see around
    obstructions
  • Communication and coordination infrastructure for
    multiple robots

Long range Link/sensors
Local sensors
robot
46
Embedded Sensors
  • Quantify impact of soil-based and air-borne
    contamination plumes
  • Monitor animal (bird) health and well-being
  • Target covers wide area
  • Selective to habitat
  • Requires robust and biocompatible chronic implants

Penetrating Neural Probe
Flexible Electrode Array
47
Towards a Unified Framework for ENS
  • General theory of massively distributed systems
    that interface with the physical world
  • Understanding and designing for the collective
  • Large-scale experiments to challenge assumptions
  • Develop key building blocks
  • Collect and share data sets
  • Develop and promulgate best-practices,
    benchmarks, design tools
  • Address emerging Ethical, Legal and Social
    Implications of pervasive monitoring

48
Pulling it all together
CENS Core Research
Academic Disciplines
Networking Communications Signal
Processing Databases Embedded Systems Controls Opt
imization … Biology Geology Biochemistry Structura
l Engineering Education Environmental Engineering
Adaptive Self-Configuration
Collaborative Signal Processing and Active
Databases
Experimental Systems
Sensor Coordinated Actuation
Environmental Microsensors
49
Enabled by…….NSF Science and Technology Center
for Embedded Networked Sensing 40Mill/10 years
base funding from NSF
UCLA Computer Science (D. Estrin (PI), R.
Muntz, S. Soatto) Electrical Engineering (J.
Judy, G. Pottie, M. Srivastava and K.
Yao) Mechanical and Aerospace Engineering (C.M.
Ho) Civil and Environmental Engineering (T.
Harmon, J. Wallace) Physiological Sciences and
Biology (P. Rundell, C. Taylor) Earth and Space
Sciences (P. Davis (PI), M. Kohler) Institute of
the Environment (R. Turco) Education and
Information (C. Borgman (PI), W. Sandoval)
Management Director D. Estrin Chief Amin
Officer Bernie Dempsey Education Coor
TBD Budget Analyst David Jaquez Admin Asst
Stacy Robinson
USC Computer Science (A. Requicha (PI), R.
Govindan, G. Sukhatme) Electrical Engineering (C.
Zhou), Marine Biology (D. Caron)
Caltech Electrical Engineering (Y. C. Tai)
UC Riverside Conservation Biology (M. Allen, M.
Hamilton (PI), J. Rotenberry)
Grades 7-12 The Buckley School (K. Griffis) New
Roads School (J.A. Wise)
Cal State LA Engineering
JPL Center for Integrated Space Microsystems
(L. Alkalai), Manager
50
Follow up
  • Embedded Everywhere A Research Agenda for
    Networked Systems of Embedded Computers, Computer
    Science and Telecommunications Board, National
    Research Council - Washington, D.C.,
    http//www.cstb.org/
  • Related projects at UCLA and USC-ISI
  • http//cens.ucla.edu
  • http//lecs.cs.ucla.edu
  • http//rfab.cs.ucla.edu
  • http//www.isi.edu/scadds
  • Many other emerging, active research programs,
    e.g.,
  • Cornell Gehrke, Wicker
  • MIT Balakrishnan, Chandrakasan, Morris
  • UCB Culler, Hellerstein, BWRC, Sensorwebs,
    CITRIS
  • UCSD Cal-IT2 activities
  • Univ. Washington Boriello
  • Wisconsin Ramanathan, Sayeed
  • DARPA Programs
  • http//dtsn.darpa.mil/ixo/sensit.asp
  • http//www.darpa.mil/ito/research/nest/

51
Conclusions
  • ENS will reveal previously unobservable phenomena
  • The network is the sensor!!
  • Building and studying real systems in the real
    (not virtual) world…in the service of real
    problems
  • Thank you to NSF, UCLA and HS-SEAS for supporting
    the vision…and to colleagues, students and staff
    for its realization.
About PowerShow.com