Sensor Web when sensor networks meet the World-Wide Web Dr. Steve Liang Assistant Professor Department of Geomatics Engineering Schulich School of Engineering University of Calgary steve.liang@ucalgary.ca http://sensorweb.geomatics.ucalgary.ca
Some background information One of the 10 Emerging Technologies That Will Change the World (2003) One of the three focused research area of GEOIDE, NCE, Canada. One of the three strategic directions of Alberta ICT. (icore)
Before Sensor Networks
Today, Sensors are everywhere.
Many Sensors are integrated with communication devices NASA Sensor Pods http://www.neptunecanada.ca MICA2 Motes
Wireless Sensor Networks become smaller and smaller smaller and smaller MIT 1984
A Macroscope and a Cyberinfrastructure With sensor networks, we are able to : Observe our physical world at very high spatial resolutions Observe our physical world continuously Collect and save the observations in digital form Some refer this concept as a Macroscope or a sensor-focus Cyberinfrastructure. E.g., Canada s Canarie NEP Programme (Network Enabled Platform)
Intelligent Instrumentation Sensor Networks are not just passive data collectors. Processing and Intelligence can be pushed into different tiers of the network At Individual sensor nodes At aggregation points within the network At Sensor network gateways At databases At Sensor Web servers At Sensor Web browsers At Users (e.g. human computing) From Passive to Active
But links between the sensor networks are missing!! Sensor Networks v.s. Sensor Web Q: What is the implication of the traffic flow monitoring to the air quality monitoring? Q: Can we re-mix these two sensor networks together for new applications??
From Sensor Networks to Sensor Web Sensor Networks Single purpose, one-time dedicated instrumentation Homogeneous platforms Device centric Proprietary and close architecture, centrally managed Specialized user interface Small and isolated Sensing It s all about the Systems Sensor Web Multi-purposes, reusability, re-mix-ability Heterogeneous platforms Network/Web enabled Open, Interoperable Many different user interfaces (e.g., mashups) Large scale, Network Effects (e.g., blogging or slogging?) Sensing and Control System, Data, and Events Modified from (Zhao, 2007)
Sensor Web Browser, a browser of the physical world Query the physical world from anywhere at anytime.
We need a World-Wide Wide SW and a SW Browser And a lot more realtime info from heterogeneous sensors. A SW Browser : Webcam Video a browser of the physical world Weather Stations Temperature Readings Traffic Sensors Spatial Data sets
Sensor Web Desires Quickly discover sensors (secure or public) that can meet my needs location, observables, quality, ability to task Obtain sensor information in a standard encoding that is understandable by me and my software Readily access sensor observations in a common manner, and in a form specific to my needs Task sensors, when possible, to meet my specific needs Subscribe to and receive alerts when a sensor measures a particular phenomenon Botts August 2004 Helping the World to Communicate Geographically
A High Level Service Oriented View of the Sensor Web 1. Standard Service Interfaces 2. Standard Information Models and Encodings
An OGC Web Service Architecture for the Sensor Web Challenge: Interoperability Methodology: Open Standard-based solution: OGC Sensor Web Service Architecture (OGC SWE) Find Publish Bind Find
Example Sensor Observation Service Client <GetCapabilities> request SOS Capabilities XML <GetObservations> request O&M XML <DescribeSensor> request SensorML <GetFeatureOfInterest> request GML Features the WWW
Examples: 1. ISIES: Intelligent Sensorweb for Integrated Earth Sensing 2. OGC OWS-3 Disaster Management Demo
ISIES - Intelligent Sensorweb for Integrated Earth Sensing A Sensor Web for a better crop yielding prediction
A Farmer s s Virtual Globe: GeoSWIFT 1.0 for ISIES Project GeoSWIFT SW Viewer GeoSWIFT Sensor Web Server
A High Level Service Oriented View of the ISIES
All data from open standard-based geospatial web services Crop yielding prediction data from a SW processing service node Background satellite imagery from a OGC Web Mapping Service In-situ sensing data from a SW in-situ sensor service node
Emergency Agency s s Virtual Globe OGC Disaster Management Demo
Disaster scene identified
Flying to the area of interest
Displaying data from different sources
Discovering weather sensors
Connecting to Sensor Observation Server
Receiving Real-time Weather Data Streams
Sensor Planning Service (SPS) for fire scanner
Tasking UAV using Sensor Planning Service request is feasible
What s s happening behind the SPS interface Image courtesy of Phillip C. Dibner, NASA Ames Research Centre
SPS sent an e-mail e to requester
SPS Results: UAV swath over Digital Raster Graphics (DRG) data over terrain
Discovering Web Cameras
Submitting Sensor Task and Receive Task ID
Sensor Observation Server for Chemical Disaster
Chemical Sensor Observation
TML Camera in Warehouse
Now, we have the standard building blocks to build an interoperable SW Through Web Service and OGC Standards We can connect heterogeneous sensor networks together. Because we re capable of connecting many sensors There will be lots and lots of sensors joining in this SW. New Challenge: Scalability What if we have millions of SW Servers??
Directory-based Sensor Web Service: MSR Sensor Map Use a similar central directory-based architecture Not Reliable: Has single point of failure Not Load Balanced: All system load added to the central registry Not Flexible: What if sensor owners don t want to publish to the company hosting the service?? (e.g., Google) BTW, Interoperability is not considered in the current design of Sensor Map architecture. Registry Client Services Microsoft Research, 2006
GeoSWIFT 2.0 High Level Architecture A P2P Architecture: Nodes leverage each others resources to <GetObservation> provide <observedproperty>urn:ssw:def:phenomenon:airtemperature</observedproperty> the service. Reliable: <featureofinterest> No single point of failure <ogc:bbox> Load <gml:envelope> Balanced: Able to distribute the system load <gml:lowercorner>38.6-77.9</gml:lowercorner> Flexibility: not controlled by a single organization <gml:uppercorner>41.0-77.85</gml:uppercorner> P2P </gml:envelope> systems have been proven scalable. </ogc:bbox> Interoperable </featureofinterest> </GetObservation> GeoSWIFT 2.0 <ObservationCollection> <observation>. </observation> <observation>. </observation>.. </ObservationCollection> System composed by peer nodes
Some projects in progress
Cyberinfrastructure and Environmental Informatics for Monitoring Biogeoscience Process in the Rockies and Foothills
Wiki-Campus (or Wiki-City?) A human participating Sensor Web + +
A stream-based publish/subscribe Spatial Sensor Web Remote Sensors Stream-based Spatial Query Processing, in-network aggregate, filtering, and compress data streams while the data moving along the path from sensors to applications. P2P Geographically-based Publish/Subscribe Architecture Enable subscription of events within areas-of-interest and real-time event notification publish Sensor Networks publish aggregate filter notify publish subscribe Sensor Web Client P2P-based Sensor Discovery Service: use a P2P-based cooperative model to discover sensing resources according to geospatial constraints. Interoperable OGC Sensor Web Standards Interoperable Web Service Interfaces and information models for the SSW
Cloud Computing for Geospatial Data and Services
Some of my thoughts about Sensor Web and Virtual Globe
The Trend of Computer Architecture: Faster, Smaller, Numerous Bell s Law (Gordon Bell, 1972) New computing class every 10 years log (people per computer) Time (Polastre, 2004)
Sensor Web as the medium between your First Earth to Second Earth If you spend more time in your second life, your first life become your first life.
What s s needed Easy to install Plug and Play Click to Install Easy to publish Think about blogging (slogging? publishing their own sensors) Easy to be discovered Ability of processing real-time data streams Alert!! Subscribe Skype, Napster, Kazza, etc.
Thank you!! Dr. Steve Liang Department of Geomatics Engineering Schulich School of Engineering University of Calgary steve.liang@ucalgary.ca http://sensorweb.geomatics.ucalgary.ca