Attendees

W-J Ketty Gann, Holger Neuhaus, Oscar Corcho, Arthur Herzog, Luis Bermudez, Michael Compton, Danh Le Phuoc, Manfred Hauswirth, Kerry Taylor, Krzsztof Janowicz, Kevin Page, Cory Henson, Amit Sheth

Introduction

Kerry on the demo we missed

Cory’s presentation

• Observation dataset
• Sensor dataset
• Linked dataset
• Sensor discovery demo (from GeoNames ‘location’)

Discussion on how+where to store observation data

• Only meta data in triples
• Values in triples - for rule based reasoning
  

- how to do that as part of the output of the SSN-Ontology work

SSN Ontology

Use of external ontologies for

• time
• location
• UoM
• Alternatives, discuss

[Ether Pad log below]

Top level concepts of the SSN Ontology: Sensor is-a Process is-a System

Semantic Markup

Presentation of draft deliverable

TODOs

• Use cases
• fill gaps in draft

Other Business

Telecon:

New timeslot for weekly meetings: Tuesday/Wednesday (alternating), 20:00 UTC

Versioning/collab. editing:

collab web Protégé: Oscar to follow up

License:

Luis to check with OGC

How do we want this to end/continue?

Output of the XG: Ontology, best practice& encoding
- is that enough for a W3C recommendation?
Option of extending the XG

--- ETHER PAD LOG ---

Semantic Sensor Network Meeting 24/10/09
Separate te physical Device from the sensor as information provider
• Device (physical) playsRoleOf sensor
• pen instanceOf Device playsRoleOf GravitySensor
• cup instanceOf Device playsRoleOf GravitySensor
• fleck instanceOf Device playsRoleOf Sensor
• Sensor = SensorInformationThing
• SensorGrounding = sort of physical thing, "the actual deployment" (find out relation to Device); May be there are three options: physical, cyber/URI, IoT
• Observation resultValue(Result)
• samplingTime
• resultTime
• Sensor outputs Result
• PhysicalPhenomena
• Process
• relation to observation and sensor
Important questions remaining:
1. how to link "interpreted" observation to transducer/pyhsical observation
2. where does process fit. System?
Sensor - Process - Observation
Alternatives
1. Ternary relation among the three of them: A sensor provides an observation using a process
(can be described with three binary relations, even so it is not exactly the same)
consequences
• In terms of OWL representation we should generate a class, since ternary relations are not directly representable.
• We give a lot of importance to the "Process" part of the representation, since it can have subclasses and/or instances
2. Attribute "uses process" in Sensor and a direct relation to observation
consequences
• The granularity of our representation takes us to the point that processes are not so important.
They are now a simple string, e.g. "Newton's law"
3. Sensors are kinds of Processes
consequences
• We have a different understanying of process not fitting well with the O&M notion
--- /ETHER PAD LOG ---

Pictures (Manfred)

Some low quality pictures taken at the F2F (phone camera):