Sensor Web Strategies

1
Sensor WebStrategies
Karen MoeSensor Web Task TeamNASA Earth
Science Technology Office February 25, 2008
CEOS WGISS-25 Sanya, China
2
Earth Observation Sensor Web

• Sensor Web Task Team (SWTT) strategies and
  expected outcomes
• Sensor web operational concepts
• Concept development since WGISS-24
• Technology push
• Technology pull
• What weve learned so far

3
Sensor Web A Service-Oriented Architecture
Approach

• Sensor webs will be dynamically organized to
• collect data,
• extract information from it,
• accept input from other sensor / forecast /
  tasking systems,
• interact with the environment based on
• what they detect or
• are tasked to perform, and
• communicate observations and results in real time.

4
SWTT Sensor Web Strategy

• Address GEOSS goals (science -gt SBA)
• Apply emerging sensor web technologies
• Leverage international resources
• EO data, models, in situ sensors, satellites
• Explore technology push pull
• Expected outcomes
• Use Cases (featuring operational concepts)
• Proof-of-concept prototypes
• Lessons learned / implications for GEOSS

5
Sensor Web Operational Concepts

• Dynamically acquire fuse data from models,
  satellite and in situ sensors
• Validate data observations in near RT
• Provide intelligent sensor control feedback to
  enable RT sensor tasking
• Enable discovery and access to sensor web
  components and services

6
SWTT Exploration Phase

• Technology push
• What sensor resources team members can bring and
  create plausible applications
• How can sensor webs support virtual
  constellations
• Technology pull
• What do scientists need to better understand and
  forecast phenomena
• What information do policy makers or disaster
  response teams need

7
Sensor Web Use Cases Explored

• Sensor web assisted Cal/Val for GRACE- CHAMP
  constellation
• Put on hold due to lack of member resources to
  pursue
• Flash flood monitoring use case builds on WGISS
  Grid technology demo
• SWTT proposed phase 1 project presentation in
  WGISS-25
• Later phases will extend demo to show model
  feedback to EO-1 sensor tasking and provide
  resulting data and forecasts to SERVIR disaster
  management system, and Intl Fed of the Red Cross
  (IFRC) global flood monitoring system

8
Sensor Web Support to ACC

• CEOS Atmospheric Composition Constellation (ACC)
  team discussed possible collaboration with SWTT
• Smoke Trajectory Forecast ACC wants to leverage
  relevant satellite and in situ sensors and evolve
  modeling approaches to overcome limitations of
  existing sensor assets to produce improved
  forecasts
• Sensor Web for ACC builds on aerosol trajectory
  model and incorporates EO-1 sensor tasking

9
Sensor Web Support to ACC
EO-1 Fire Sensor Web Evolution

• CALIPSO (near RT aerosol data) and MODIS
  (vertical component data) augment model forecast
• Produce a 3D smoke trajectory forecast product to
  international AirNow system
• Compare predicted with actual smoke conditions
  using EO-1 imagery

MODIS Active Fire Map
EO-1 (ALI Hyperion)
Smoke Trajectory Forecast Model
10
Underlying Sensor Web flow

• Example sensor web themes in use cases, an
  emerging pattern
• Routine event monitoring (in situ rain gauges,
  sentinel systems for fire, volcanos, etc)
• Model predicts potential event (flood, smoke
  trajectory)
• Event detection or model prediction triggers
  request for near RT sensor observation task
• New observation data augments model for more
  accurate forecast
• New observation and improved forecast feeds
  disaster management portal

11
Reflections on SWTT Activities since WGISS-24

• Lessons learned on how we work as a team in
  support of CEOS
• Discuss need for documenting SSWT
• Program perspective
• Project perspective
• Use Case
• Activity Flow Chart
• Findings and Recommendations

12
Reflections on SWTT (contd.)

• Program perspective
• A strategic view of the team activities
• One over-arching profile of team, expected
  outcomes, relating activities to GEOSS
• Project perspective
• A short project plan describing the objectives of
  a selected application prototype
• One each per application flood, wildfire/smoke

13
Reflections on SWTT (contd.)

• Use Case
• A detailed discussion of a specific application
  summarizing actions, actors, resources
• Activity Flow Chart
• A very detailed diagram of the source and sink of
  each step of the prototype demo
• Methodology presented by M. Burnett
• An Approach for Repeatable Sensor Web
  Construction

14
Reflections on SWTT (contd.)

• Findings and Recommendations
• Summarize our findings on what worked, what
  didnt work, other approaches to try
• Describe experience (pros, cons) with
• Standards
• Processes
• Tools
• Make recommendations
• CEOS
• GEOSS
• Standards bodies (ISO, OGC, others)

15
Back-up Charts

• Overview of use case progression since WGISS-24
• WGISS-24 sensor web technology challenges

16
Sensor Web Use Cases Explored

• Sensor web assisted Cal/Val for GRACE- CHAMP
  constellation, involving taskable weather
  balloons, associate with GPS water vapor profiles
• CHAMP-GRACE constellation used to profile water
  vapor
• Task in situ weather balloons
• Implement web services to discover and task
  applicable weather balloons
• Fuse data products and identify mismatches where
  calibration is needed

17
Building on Grid Technology Demo

• Flash flood monitoring Rain gauge input to
  forecast model detects potential flood condition.
  Improve flood model forecasts by discovering and
  supplying recently acquired applicable satellite
  data
• Rain gauge sensors on Zambezi River (Mozambique
  seasonal flood)
• NASA data identified via ECHO services
• NASU model forecasts flood conditions

18
Model Feedback to Sensor Tasking

• Flash flood monitoring Model flood forecast
  triggers EO-1 tasking event. Resulting image
  delivered directly to first responders
• NASU model forecasts flood triggers EO-1
• EO-1 acquires current image
• Model forecast accuracy is improved
• Satellite image also delivered directly to
  SERVIR, a disaster response system initially
  developed for Central and South America, now
  being applied to African events

19
Sensor Web Extension to IFRC

• Flash flood monitoring Intl Fed of Red Cross
  Red Crescent (IFRC) approached NASA about
  incorporating satellite data to improve existing
  and planned global flood monitoring of 200 sites
  world wide
• Team from Geneva is providing operational user
  insight to use case
• IFRC has disaster response planning system and
  staff interested in improved information
  available from use of NASU model and RT sensor
  tasking in EO-1

20
Sensor Web Features and Benefits

• Some Features
• Targeted observations through dynamic tasking
• Incorporate feedback to adapt autonomous
  operations (e.g., weather forecasts)
• Ready access to data and information
• Some Benefits
• Improved resource use and reuse through
  reconfiguration of assets
• Improved cost effectiveness through autonomous
  operations
• Rapid response to evolving, transient phenomena
• Improved data quality and science value by
  comparing sensor data from the same event

Derived from the NASA ESTO Sensor Web Meeting Feb
2007
21
1. Technical Challenges

• In the collection and analysis of information
  from heterogeneous nodes
• There is a lack of uniform operations and
  standard representation for sensor data
• There exists inadequate means for resource
  reallocation and resource sharing
• Deployment and usage of resources is usually
  tightly coupled with the specific location,
  application, and devices employed

22
2. Technical Challenges

• Publishing and discovering sensor resources
• Create a publicly accessible infrastructure for
  publishing heterogeneous sensor resources and
  complex applications
• Discover and use sensor resources
• Sensor data fusion
• Sensor data has different data models and formats
  and different spatial and temporal resolutions,
• Fusion -gt higher spatial coverage and temporal
  resolution

23
3. Technical Challenges

• Context-based information extraction
• End users have insufficient technical expertise
  and time to extract information from sensor data
• Users require different views of the data
  according to needs and context
• Data can be filtered, summarized, transformed
• Features can be extracted -gt higher level
  features -gt information -gt application/decision
  making
• Same data can be reused for different
  applications

24
WGISS Sensor Web Discussion

• Identify Collaboration Opportunity
• Standards-based proof-of-concept sensor web demo
• Applied to significant GEO objective (e.g.,
  Virtual Constellation?) identify GEO champion
  user(s)
• Mature standards, capture lessons learned
• Develop processes, toolkits to improve usability
• Leverage NASA Earth Science Sensor Web technology
  investments and prototypes
• Provide feedback to standards bodies, e.g.
• OGC SensorML, Mike Botts/UAH
• OGC SWE, Liping Di/GMU, Stefan Falke/NG, others
• Other standards?
• Formally recommend proven standards to GEOSS