Deborah Estrin

1
CENS Some highlights from our first 1.5 years

• Deborah Estrin
• Work summarized here is largely that of students,
  staff, and other faculty at CENS--see poster
  session and handouts for detailed attribution
• We gratefully acknowledge the support of our
  sponsors, including the National Science
  Foundation, Intel Corporation, Sun Inc., Crossbow
  Inc., and the participating campuses.

2
Mission Statement

• To address scientific issues of national and
  global priority through pioneering research and
  education in Embedded Networked Sensing
  technology.
• To develop and demonstrate architectural
  principles and methodologies for deeply embedded,
  massively distributed, sensor-rich distributed
  systems
• To apply and disseminate these systems in support
  of scientific research critical to social and
  environmental concerns
• To create meaningful inquiry-based science
  instruction using embedded networked sensing
  technology, for a diverse grade 7-12 population
  and to disseminate education materials and
  technology through outreach and professional
  development networks

3
Embedded Networked Sensing

• 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

Contaminant Transport
Ecosystems, Biocomplexity
Marine Microorganisms
Seismic Structure Response
4
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, Atmel, TinyOS
Medusa, MK-2 UCLA NESL 2002
Telos Mote UCB, 2004 Zigbee radio, Motorolla
Energy is the primary resource constraint and
communications is the primary consumer
5
Technology Design Themes

• Long-lived systems that can be untethered
  (wireless) and unattended
• Communication will be the persistent primary
  consumer of scarce energy resources (Mote
  720nJ/bit xmit, 4nJ/op)
• Autonomy and highly dynamic, irregular
  environments requires robust, adaptive,
  self-configuring systems
• Leverage data processing inside the network
• Exploit computation near data to reduce
  communication, achieve scalability
• Collaborative signal processing and localized
  algorithms
• Flexible tasking incorporating models, analysis,
  fusion with other data sources
• The network is the sensor (MangesSmith,
  Oakridge Natl Labs, 10/98)
• Requires robust distributed systems of hundreds
  of physically-embedded, unattended, and often
  untethered, devices.

6
Overview

• Information Theoretic Foundations
• Technology
• Systems
• Signal processing
• Actuation
• Sensors
• Applications
• Habitat Sensing
• Contaminant Transport
• Marine Microorganisms
• Seismic Monitoring
• Education

7
Information Theoretic Foundations
8
Scalability for Point Sources in Sensor Networks
Scalability for Point Sources in Sensor Networks

• Information theory concerned with fundamental
  limits
• Capacity maximum reliable communication rate,
  versus bandwidth and power
• Rate-Distortion minimum required rate to
  describe source, subject to distortion constraint
  
• Gupta Kumar early result showed wireless
  communication networks do not scale with node
  density per node capacity goes to zero
• However Cooperative rate distortion coding
  results in most communication being local in
  sensor networks-- more nodes do not necessarily
  result in more traffic
• More relays enable increased frequency re-use
  capacity can increase without bound

9
Scalability for Distributed Sources
Scalability for Distributed Sources

• To estimate parameters of a field (e.g., to get
  isotherm map) information increases until achieve
  desired spatial sampling
• After this extra nodes contribute no additional
  information, but can increase communication
  resource
• Image processing analogy specify pixel size
• Parameters to describe local field can be compact
  compared to raw data, for given level of
  distortion

10
Practical Implementation
Practical Implementation

• Dense network in neighborhood have mix of nodes
  with different ranges, operating in separate
  bands
• Perform local fusion to avoid long-range
  communication of raw data
• Locally route towards the longer range links
  they act as traffic attractors, causing number of
  hops at any given layer to be small, limiting
  delay
• Result is a (largely) standard overlay
  hierarchical network

11
Technology

• Systems
• Services, Macroprogramming, Tools
• Signal Processing
• Actuation
• Sensors

12
Technology Systems
13
Common Services/Tools for Robust, Scalable,
Flexible, Deployable Systems
Common Services/Tools for Robust, Scalable,
Flexible, Deployable Systems
Localization Time Synchronization
Calibration
In Network Processing
Programming Model
Routing and Transport
Event Detection

• Needed Reusable, Modular, Flexible,
  Well-characterized Services/Tools
• Time synchronization, Localization, Calibration,
  Energy Harvesting
• Routing and transport
• In Network Storage, Querying, Processing, Tasking
• Macro-Programming

14
Localization of Sensor Nodes

• Robust ranging
• Wideband acoustics
• Scalable distributed algorithms
• Collaborative multilateration(with beacons)
• Geometry-driven beacon-less
• Fundamental error analysis
• Cramer-Rao bounds for multihop
• Geometry effects
• Angle vs. distance
• Implementation
• MK-II platform with ultrasound ranging

15
Energy Harvesting
Energy Harvesting

• Need distributed methods to learn the
  environmental energy opportunity at all nodes
• Global task sharing among nodes to optimize
  performance

• HelioMote test-bed
• Recharge batteries from solar
• Track energy received
• Monitor residual battery status
• Provide constant voltage to load as battery
  voltage degrades

NIMS aerial nodes also solar powered, self
sustaining
Environmentally aware
Battery based
16
Directed Diffusion Adaptive Data-Centric
Routing and Transport
Directed Diffusion Adaptive Data-Centric
Routing and Transport
Sink

• Data-Centric routing supports in network
  processing
• Tinydiffusion implements one-phase-pull variant
• Pulls data out to only one sink at a time (saves
  energy)

Sources
Interest
Routed Data
Gradient
17
Characteristics of Wireless Connectivity
Asymmetry vs. Power
Reception v. Distance
What Robert Poor (Ember) calls The good, the
bad and the ugly
Standard Deviation v. Reception rate
18
Scalable Infrastructure for Data Storage

• Goal
• Build a flexible event storage systems for sensor
  networks
• Components
• Networking primitives
• Distributed data structures such as hash tables
  and multi-dimensional indices
• Query optimization
• Supports different programming abstractions
• Database/declarative
• Tuple-space
• Logic programming
• Flexible
• Components assembled at compile time

19
Dimensions Lossy Multi-resolution Data Aging in
Storage-constrained Networks

• Goal
• Building a long-term in-network storage
  infrastructure for storage-constrained networks.
  Exploit spatio-temporal correlation in sensor
  data, distributed storage capacity and training
  datasets to achieve goal.

30

• Key Ideas
• Construct lossy wavelet-compressed summaries
  corresponding to different resolutions and
  spatio-temporal scales.
• Queries drill-down from root of hierarchy to
  focus search on small portions of the network.
• Progressively age summaries for long-term storage
  and graceful degradation of query quality over
  time. Use training data to determine aging
  periods.

Example Query Find nodes along a boundary
between high and low precipitation areas.
Error
5
Only coarsest summary is queried.
All resolutions (coarsest to finest) are queried
20
Macroprogramming

• State of the art
• Components, initializations, and wirings are
  handcrafted for applications optimized by eye,
  not by a compiler
• Complexity has required application development
  by systems programmers, not end users
• Objective
• Automate the difficult parts of application
  construction. Its goal is to provide a way for
  non-programmers to easily develop efficient
  sensor network applications

21
Sensor Network Application Construction

• Problem
• Component model of application development is
  inherently hierarchical functional redundancy
  leads to structural inefficiency
• Macroprogramming Solutions
• Users describe service requirements, not specific
  wirings
• Compiler weaves together underlying components to
  optimize structure
• Compiler merges components with compatible
  initializations
• Compiler chooses components with superset
  functionality

22
Em Development and Deployment Software
Collaborative Sensor Processing Application
Domain Knowledge
3d Multi- Lateration
State Sync
Reusable Software
(Flexible Interconnects not a strict stack)
Topology Discovery
Acoustic Ranging
Neighbor Discovery
Reliable Unicast
Leader Election
Time Sync
Radio
Sensors
Audio
Hardware
23
Connectivity Measurement Tool
SCALE Connectivity measurement system based on
Emstar
Motes used to transmit and receive packets -- A
real-world augmentation to a virtual simulation
24
Technology Signal Processing
25
Collaborative Signal Processing
Real-time Acoustic Localization using AML
Algorithm
Bearing crossing of 2 DOAs yields locations
DOA estimation

Source
Source
iPAQ
iPAQ
Fine-grain Time-Synch. Wireless LinkedTestbed
One subarray (4 iPAQs)Two subarrays (4 iPAQs
each)

• Reverberant DOA/Localization
• Most DOA/localization algorithms will not work in
  moderate-strong reverberant environments
• Source image model reflects source along all
  relevant reflecting walls--Only useful for
  simulation
• New array image model reflects array along
  relevant reflecting walls--useful for actual
  DOA/localization
• Simulation of circular subarray (8 iPAQs) with
  strong reflecting left and bottom walls for DOA
  estimation

26
Optimum Sensor Placement

• Cramer-Rao Bounding Approach
• CRB of source location error variance
• For 8 fixed sensors, error variance is low when
  source is inside convex hull of the sensors

• Minimum Entropy Approach
• Given locations and sensing certainties,
  Bayesian method can compute target location
  distribution under minimum entropy criterion
• Bayesian bounds equal CRBs when sensing
  uncertainties are Gaussian

Smaller error variance inside convex hull of 4
sensors
Larger error variance outside convex hull of 4
sensors
27
Technology Actuation
28
Actuation Networked Info-Mechanical Systems
(NIMS)
Sensor-Coordinated Mobility
Actuation Networked Info-Mechanical Systems
(NIMS)

• NIMS Architecture Robotic, aerial access to full
  3-D environment
• Enable sample acquisition
• Coordinated Mobility
• Enables self-awareness of Sensing Uncertainty
• Sensor Diversity
• Diversity in sensing resources, locations,
  perspectives, topologies
• Enable reconfiguration to reduce uncertainty and
  calibrate
• NIMS Infrastructure
• Enables speed, efficiency
• Provides energy transport for sustainable presence

29
Actuation Fundamental NIMS Algorithms

• Adaptive Sampling
• Resource-efficient environment mapping
• Simulation and experimental characterization
• Sensing Diversity
• Reduction of sensing uncertainty w/ constrained
  mobility
• Theoretical analysis and experimental mapping of
  obstacle-field environment by distributed imaging
• Coordinated Mobility
• Multi-node scheduling

30
Adaptive Sampling for Environmental Robotics

• Robot as Geostatitics agent
• Creating a dynamic Map of the environment

• Divide and Conquer
• Stratify the current cell into four
• Collect data in current cells
• Calculate the variance
• Iterate until variance is below threshold

31
Actuation NIMS Systems

• NIMS System Development
• NIMS Field Systems
• Laboratory NIMS installations enabling
  development, verification
• NIMS Angle-Resolved Imaging Spectroscopy
• NIMS Metrology (NIMS node geolocation)
• Emstar systems support
• NIMS Deployments
• Wind River Canopy Crane Research Facility
• 50m x 50m transect (9/03)
• James Reserve
• 70m x 15m transect (1/04)

32
Actuation Experimental Testbeds
An underwater robotic network for marine
micro-organism detection
Autonomous deployment of a sensor network
from an aerial robot
Pioneer mobile robots deploying, repairing and
navigating using a sensor network
The Robomote
33
Robot Navigation using a Sensor Network

• Done at Intel in late summer 2003
• Mica2 mote-based sensor network
• Mobile robot navigates based solely on network
  directives
• Results include over one km of robot traverses in
  experiments, and an Intel patent filing

Sensor node
Robot
34
Network Deployment and Repair from the Air

• Mica2 mote-based sensor network, Mobile robot is
  an autonomous helicopter
• Results include network deployment and repair
• Significant external collaboration (D. Rus, MIT,
  P. Corke, CSIRO, Australia)

Initial Deployment
After Repair
35
An Actuated Underwater Sensor Network for Marine
Microorganism Monitoring

• Mica2 mote-based sensor network, Mobile robot is
  an autonomous submarine
• Results
• Binary search algorithm for approximating
  thermocline
• Using the submarine as a data mule to prolong
  network lifetime

36
RoboGaming
Graphic Projector
Localization Camera
Graphical play authoring tool
PicoNIMS Terrain Control
Energy Replenishment
Autonomous Mobile agents
Game Server
Reconfigurable Terrain
Distributed Audio Engine

• Real agent motion beyond computer graphics
• Play with autonomous robots and reconfigurable
  structures
• Physical capabilities and constraints replace
  simulated effects
• game more interesting

37
Technology Sensors
38
Development of New Embedded Sensors

• CENS-compatible chemical-sensor technologies
  needed
• soil/water quality monitoring, security,
  precision agriculture
• Current chemical-sensor development
• electrochemical nitrate
  sensorspotentiometric orchronocoulometry
• MEMS liquid chromatography systems and mass
  spectrometers

39
Nitrate Sensor Development
Contaminant Transport Group
short term medium term long term

• Potentiometric nitrate sensor Det. Lim. ppm
• amperometric nitrate sensor Det. Lim. ppb
• LC-on-a-chip separation and identification of
  ions
• surface plasmon resonance ultra-sensitive

40
Applications

• Habitat Sensing
• Contaminant Transport
• Marine Microorganisms
• Seismic Monitoring

41
Science Application System Development

• Biology/Biocomplexity
• Microclimate monitoring
• Triggered image capture
• Contaminant Transport
• County of Los Angeles Sanitation Districts
  (CLASD) wastewater recycling project, Palmdale,
  CA
• Seismic monitoring
• 50 node ad hoc, wireless, multi-hop seismic
  network
• Structure response in USGS-instrumented Factor
  Building
• Marine microorganisms
• Detection of a harmful alga
• Experimental testbed w/autonously adapting sensor
  location

42
Application Habitat Sensing
43
Ecophysiological Modeling Using Sensor Array Data

• Spatially and temporally dense microclimate data
  will allow
• significant advancements in modeling plant
  production

44
Habitat and Environmental Sensing Applications
UC James Reserve Habitat Sensing Testbed
NIMS mobile ground and canopy climate sensors,
data mules, and robotic samplers
Dense micro-climate sensor networks Extensible
Sensor System (ESS)
Cavity nest micro-climate, remote observation,
bioacoustic sensing
Soil microclimate and chemical sensors,
root/fungi imaging systems (mini-rhizotron)
James Reserve and Hall Canyon Research Natural
Area
45
System Support for HabitatMonitoring Extensible
Sensing System (ESS) Software

• Tiered architecture to support taskable, scalable
  monitoring
• Mica2 motes (8 bit microcontrollers w/TOS) with
  Sensor Interface Board hosting in situ sensors
• Tinydiffusion provides tasking and multihop
  transport over SMAC links
• Microservers are solar powered, 32-bit
  processors, linux OS
• Pub/sub bus over 802.11 to Databases,
  visualization and analysis tools, GUI/Web
  interfaces

46
System Support for Habitat Monitoring Sensor
Interface Board

• A general framework for attaching multiple
  instances and different types of sensors
• ADC true 12bit
• High gain differential channels
• Digital Input (Interrupt driven)
• Digital Outputs
• Counter, frequency
• Relay output
• On board voltage,temperature and humidity
• Flexible sampling rate
• Configurable for different translation functions
  per channel based on the sensors that has been
  attached
• Tested with different sensor types.
• http//www.cens.ucla.edu/mhr/daq/
• Now sold by Crossbow

47
Application Contaminant Transport
48
Error Resilient Contaminant Monitoring

• Sensor network error resiliency in complex media
  (air-water-soil)
• Working in the context of a real problem in
  Palmdale, CA
• partnering with LA County Sanitation District
• Working in concert with Sensor Group on broadly
  applicable sensors, scalable sensors nitrate and
  other ionic species
• microsensors matching COTS perfomance
• Real-time analysis instead of logging
• model calibration, forecasting

49
Contaminant Transport Futures

• Larger scale, multimedia problems
• Linking remote and in situ sensing over multiple
  scales
• Management, visualization, exploration of
  massive, heterogeneous data streams
• NSF CLEANER Initiative

50
Data Models for Environmental Monitoring
Contaminant Transport Group

• Challenges
• Multimedia, Multiscale problems (time and space)
• Multidisciplinary (current and as yet unknown)
  problems
• Management, visualization, exploration of
  massive, heterogeneous data streams
• Few standards and discipline/problem specific
• Eecology specific Ecological Metadata Language
  (EML), Content Standards for Digital Geospatial
• Cross-cutting standards for CENS application
  SensorML, FGDC - Geographic Information Systems
  (GIS)
• Competing metadata standards exist for
  educational objects--descrobe scroted activities
  for data use, not descriptive information about
  data themselves
• IEEE Learning Object Metadata (LOM)
• The Gateway to Education Materials (GEM)
• ADEPT/DLESE/NASA metadata (AND)

51
Application Marine Microorganisms
52
Detection of Marine Micro-organisms using
Immuno-based Methods
Experimental Testbed
AFM Image of Alga

• Detection of a harmful alga (Brown Tide ) using
  AFM and ELISA and Flow Cytometry
• Experimental testbed Glass columnw/ sharp
  thermocline and autonously adapt sensor location
  to thermocline
• Prptocol for counting using the FITC-conjugated
  MAb
• Cell imaging w/ AFM on poly-L-lysine coated mica
  surfaces

Flow Cytometry
53
Marine Micro-Organism Detection Experiment
Temperature profile and growth of Brown Tide alga
with depth in column
Brown Tide Cells/ml
thermocline
0 2x106 4x106
6x106
T1 day
0
T3 day
T7 day
T13 day
50
T15 day
T18 day
T21 day
Depth (cm)
T22 day
100
T24 day
T27 day
150
Addition of BT grazer, Pedinella
200
0 5 10 15 20
25 30
Temperature (C)
54
Marine Micro-organism Detection Experiment
Growth of Brown Tide alga at depths of 2-12cm and
134-144cm in column
Addition of BT grazer, Pedinella
6x106
4x106
Brown Tide Cells/ml
2x106
o
0 5 10 15 20
25 30
Time (days)
55
Application Seismic Monitoring
56
Seismic Applications

• Multi-Hopped Radio Linked Array features
• Time synchronization
• Network event detect
• Sequenced event transmission
• Deployments planned for UCLA campus and the San
  Andreas Fault (100m-10 km)
• Easily reconfigurable
• Worldwide application
• Factor Building site
• 72 channels of 24-bit data
• 500 samples per second continuous data recording
• Internet accessible real time data monitoring
• Observation of 4 strong earthquakes, including
  Alaska Japan

Fiberoptic link
Radio link
57
Factor Building Motion

• Spectral characterization of building
• Better detection of damage following earthquakes
• Track long term changes in building and soil
  strength
• Recognize effects of environment (wind, rain,
  etc.)

Fundamental Mode
1st Overtone
58
Hardware Plans for Seismic Deployment
Stargate (Intel, Crossbow) X-Scale based,
400MHz 32MB RAM, Flash runs Linux
Kinemetrics digitizer
Guralp Seismometer
802.11 fast back-channel, command and time
distribution
59
Systems Support for Structural Monitoring

• Goal
• Develop system capabilities needed to support
  dense structural monitoring
• Projects
• Wireless mote-based array for Factor and
  Four-Seasons experiment
• Sensor/actuator array for periodic localized
  testing of structures for damage
• GPS-less Timesynch for wireless sensor arrays

60
New Directions
Security
Theatre, Film, Television
Precision Agriculture
Tropical Biology
Coral Reef
Gaming
Global Seismic Grids/Facilities
61
Education
62
Grade 7-12 Science EducationSensor Networks as
Experimental tool
63
Undergraduate Research Experience

• Undergraduates exploring new interdisciplinary
  domains

• Originally, I really only considered a masters
  program, but seeing the doctoral program and
  hearing about it from lab members definitely
  generated greater interest on my part to engage
  in a similar atmosphere in the future.Undergradu
  ate Scholar
• If mentoring an intern makes up their mind to
  attend graduate school or increase their
  interest, then mentoring is worth it.Graduate
  Student Mentor
• The CENS program went beyond the expected gave
  me more learning experiences than anticipated.
  Undergraduate Scholar

• Exploratory work for new grants
• Teamwork training
• Increased enthusiasm
• Mentor system of grad students
• 45 undergraduates participated summer 03
• 26 were women
• 26 were minorities

64
Ethical, Legal, Social Implications
Pervasive Computing The ethical, legal, and
policy issues must be addressed during the design
and use stages of these Embedded Network
systemsA more in-depth analysis of public
policy issues is urgently needed that would lead
to appropriate recommendations for solving likely
problems.National Academy of Sciences

• Interesting Developments
• RFIDs You might not care about someone
  tracking your razor bladesbut what about your
  tires?
• Camera phones
• Fusion of sensor modalities

65
Conclusions

• Center concept allowed opportunities to emerge
  and be realized at a rate that exceeded our
  expectation
• Systems, signal processing and actuation have so
  much in common its difficult to maintain separate
  web pages!
• Applications are sharing platforms and building
  blocks
• Less-GPS Timesynch for Seismic and acoustic
  arrays
• Contaminant and Ecosystem soil array platform
  hardware and software
• One size does not fit all, nor perhaps any all
  of our applications will be served by tiered
  architectures
• The Robotic Ecology vision is becoming real
  through NIMS
• Undergraduate research proved itself as highly
  effective and mutually beneficial to the students
  and CENS
• Some challenges to early deployments were even
  more interesting than we expected
• Building deployable systems with resource
  constrained devices and aggressive duty cycling
  is hard!
• In situ calibration, localization dont yet have
  deployable early incarnations

66
Background and 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/
• Conferences ACM Sensys (Nov 03), WSNA, IPSN,
  SNPA (ICC), Mobihoc, Mobicom, Mobisys, Sigcomm,
  Infocom, SOSP, OSDI, ASPLOS, ICASSP,
• CENS website http//cens.ucla.edu (posters from
  recent research review)
• Whose involved
• Active research programs in many CS (networking,
  databases, systems, theory, languages) and EE
  (low power, signal processing, comm, information
  theory) departments
• Industrial research activities at Intel, PARC,
  Sun, HP, Agilent, Motorola
• Startup activity at Crossbow, Sensicast, Dust
  Inc, Ember,