Semantic Representation of Sensor Networks Data

1
Semantic Representation of Sensor Networks Data
Presented By Eman Ibrahim (eibrahim_at_site.uottawa
.ca) Mohamad Eid (eid_at_mcrlab.uottawa.ca)
ELG 7871B Networked Appliances, Home Networking,
and Pervasive Computing, Service Discovery Dr.
Ramiro Liscano Nov. 29, 2005
2
Outline

• Introduction to Sensor Networks
• IEEE 1451 Standard for Sensor Networks
• Overview of TEDS and Templates
• Drivers for Semantic Sensor Data
• Ontology Definition
• Advantage of Semantic Representation of Sensor
  Data
• Taxonomy for Sensor Ontology
• Implementation and Analysis
• Demonstration
• Recommended Future Work
• Questions

3
Introduction Sensor Networks

• Sensor physical device that detects a signal or
  physical condition
• Gateway is a device that supports a network
  interface, application functionality, and access
  to the sensors data on the Internet.

4
IEEE 1451 Standard for Sensor Networks

• Developed to standardize transducer interface
• 1451 Standard Objectives
• Common communication interface between
  transducers
• Compatibility with multiple sensor bus standards
• Interconnect analog transducer with digital
  networks
• Not developing a new network standard
• Expected Advantages
• Allow interoperability of transducers of
  different vendors
• Allow the use of existing control system
  installation
• Allow transducers to share a common bus
• Increase the usage of existing networks

5
IEEE 1451 Family of Smart Transducers Interfaces
6
Transducer Electronic Data Sheets (TEDS)

• Contains technical information that identifies
  the sensor, specifies the sensors analog
  interface, and describes the sensors use.
• TEDS resides in the sensor in an inexpensive
  memory component, typically an EEPROM.
• Advantages
• Eliminates the need to manually input data when
  configuring a system or sensor
• Only uses 256 bits of memory
• Electronics are inexpensive memory components
• Typically uses electrically erasable programmable
  read-only memory which communicates digitally to
  the data acquisition system

7
Standard TEDS Contents
8
Components of TEDS

• Basic TEDS (first 64 bits of the transducer TEDS)
  contains basic Identification information such
  as
• Manufacturer ID ( 14 bits)
• Model Number (15 bits)
• Version Letter (5 bits) and Number (6 bits),
• Serial Number (24 bits) of Transducer
• IEEE Standard TEDS contain the technical
  information for the sensor behavior of
  transducer
• Measurement range,
• Electrical output range,
• Sensitivity (at reference frequency),
• Power requirements, and
• Calibration information (i.e. last calibration
  date)

9
Standard Templates-Transducer Types
10
Thermocouple Template
11
Why Semantics for Sensor Networks?
12
Advantages of Semantic Sensors

• Improving Search
• The search engine has info about the meaning of
  terms.
• Increases the precision rate
• Decreased the recall rate
• Information Access
• Enable scalable sensor information access
• Semantic services
• Semantic description of sensor data facilitates
  semantic web services to process and reason
  sensor data

13
Ontology Definition

• An explicit formal specification of a shared
  conceptualization
• Explicit concepts and constraints are explicitly
  defined
• Formal machine readable
• Shared captures consensual knowledge
• Conceptualization abstract model of phenomenon
• The Ontology Development Life Cycle
• Obtaining an Initial Vocabulary List
• Identifying an Initial Taxonomy
• Adding Restrictions and Axioms
• Consistency Checking
• Incremental Modifications
• Evaluation

14
Initial Taxonomy for Sensor Ontology
Curve
Calibration
Frequency_Response
Table
Data
Physical Unit
Format
Prototype
THING
Electrical
Parameters
Location
Actuator
Identity
Operator
Sensor
Manufacturer
Owner
Transducer
Physical
15
Implementation Tools

• Protégé
• Free, open source, Java based Ontology Editor
• Knowledge base framework
• Supports Frames, XML schema, RDF(S), and OWL
• See http//protege.stanford.edu/
• RacerPro (RACER)
• Middleware for the Semantic Web
• Offers reasoning services
• Compute the classification hierarchy
• Check the logical consistency
• http//www.racer-systems.com/

16
Components of OWL ontology

• Classes
• Concrete representation of concepts
• Sets that contain individuals
• Examples transducer, sensor, owner, manufacturer
• Properties
• Are relationships on individuals
• Examples can_Be, of_Type, characterized_By
• Several types
• Functional Properties
• Inverse Functional Properties
• Transitive Properties
• Symmetric Properties
• Domain and Range
• Individuals
• Instances of classes
• Example MICROCIOL is an instance of Sensor class

17
Building the Ontology - Procedure

• Create an OWL files project
• Create classes Data, Sensor, and all subclasses
  (taxonomy)
• Create properties
• Object Properties among individuals
• Prototype Properties between individual and
  schema data type value
• Annotation Properties metadata to classes,
  properties, or individuals
• Describing and defining Classes
• Property Restrictions existential and universal
• Create instances of classes and bind them using
  properties

18
Testing the Ontology

• Two main tests
• Subsumption Testing
• Test whether or not a class is subclass of
  another
• Inferred ontology class hierarchy is computed
• Compared to the asserted hierarchy
• Consistency Check
• Based on the description/conditions of a class
• Checks if a class can have any instances
• A class is deemed inconsistent if it cant have
  any instance
• Incremental Modification
• Ongoing process of growing classes and individuals

19
Racer Reasoning with OWL

• Using RACER
• Support multiple ontologies
• Standalone server versions available for Linux
  and Win
• Network based API supported (HTTP, TCP/IP)
• The only true reasoner for individuals
• Invoking the Reasoner
• To automatically classify the ontology Classify
  Taxonomy
• To check the consistency, the Check Consistency
  should be used
• The reasoner computes the inferred hierarchy
• A class can be reclassified if its super class
  changes (marked blue)
• An inconsistent class will be circled in red

20
Inconsistency Check Example
Manually constructed
Generated by the reasoner
21
Demo Knowledge Representation

• How would one create the following knowledge
• The voltage has a unit of volts
• The voltage has float prototype
• Float prototype is defined by
• Number of bits Eight
• Start Value E-3
• Tolerance E-5
• Lets take some time to explore

22
Recommended Future Work

• Complete the taxonomy to describe different
  classes of sensors
• Developing comprehensive vocabulary list (tokens)
  using software
• Define a functional ontology that describes
  operations on data

23
THANK YOU