Sensor Web

1
Sensor Web

• Browsing the physical world in real-time
• By Vincent Tao

2
History

• Conceived 1997, NASA/Jet Propulsion Laboratory

3
History

• Conceived 1997, NASA/Jet Propulsion Laboratory
• Idea
• sensor chips to monitor and control environment
• sensations transmitted via internet in real-time

4
History

• Conceived 1997, NASA/Jet Propulsion Laboratory
• Idea
• sensor chips to monitor and control environment
• sensations transmitted via internet in real-time
• Why is this different?

5
History

• Conceived 1997, NASA/Jet Propulsion Laboratory
• Idea
• sensor chips to monitor and control environment
• sensations transmitted via internet in real-time
• Why is this different?
• Cheaper sensors ? more possibility
• Networked (not individual) sensors

6
Example Restaurant Waiting Time
7
Problems Interoperable

• Interoperable
• In-site (on-site) sensors
• Measuring physical properties of an area
• Remote sensing
• Via radiation reflected or emitted from object
• GPS
• Web

8
Problems Interoperable

• Interoperable
• In-site (on-site) sensors
• Measuring physical properties of an area
• Remote sensing
• Via radiation reflected or emitted from object
• GPS
• Web
• Implies data formatting standards (Open GIS)
• Web Map Service, Geographic Markup Language,
  SensorML, etc

9
(No Transcript)
10
Problems Intelligent

• Intelligent
• Sense environment and respond
• Communicate amongst each other
• Data integration

11
Problems Flexible

• Flexible
• Different types of sensors
• Deterministic
• Triggered
• On-demand
• Broken sensors
• Fault tolerant to noise, redundant
• Weather conditions
• Easy deployment

12
Problems Scalability Size

• Scalability
• Large number of interacting sensors
• Large number of simultaneous requests
• Efficiently locate sensors
• Adapt to sensor join/leaves
• Size does one size fit all?

13
Problems Scalability Size

• Scalability
• Size does one size fit all?
• smaller panels (less energy harvesting)
• smaller antennae (less radio range)
• smaller batteries (less power)
• larger sensors (harder to manage, intrusive)
• ? What about tiered architecture?

14
SenseWeb

• Microsoft Research project
• Goals
• Ease of data publication
• Application-to-application compatibility
• Primitives to query live sensors
• ? Create SensorNet
• Display MSN Virtual Earth

15
(No Transcript)
16
SenseWeb Architecture

• Data Publishing Toolkit (DPT)
• Publishes sensor metadata (location/type) to
  GeoDB
• Publishes sensor data in response to queries
• Uses sensor ontology standards
• GeoDB
• Indexes metadata uses hierarchical triangular
  mesh (HTM)
• Queries submitted by keyword, location, etc.

17
SenseWeb Architecture

• Data Publishing Toolkit
• GeoDB
• Aggregator
• Data integration mashes sensor data with client
  side GUI (e.g. MSN Virtual Earth overlay)
• User query ? GeoDB relevant sensors ? DPT
  real-time data ? aggregate/summarize

18
SenseWeb Architecture

• Advantages
• Data owner only talks to DPT
• End users only browse web page and query data
• ? GeoDB and Aggregator transparency
• Disadvantages?

19
Conclusion

• Ubiquity Invisibility
• Overview of problems
• Four layers
• Sensor sensor design, materials, etc
• Communication protocols, standards, etc
• Location routing, addressing, etc
• Information data integration, distribution, etc
• Which can be different? Which need
  standardization?

20
Conclusion

• Next algorithms techniques to solve problems
  (e.g. HTM, P2P communication, statistical data
  modeling, more applications)