Configuration Strategies of an AR Toolkit

Configuration Strategies of an AR
Toolkit-based Wide Area Tracker
The Second IEEE International
AR Toolkit Workshop
Martin Wagner, Felix Loew
Lehrstuhl für Angewandte Softwaretechnik
Institut für Informatik
Technische Universität München
wagnerm@in.tum.de
16-Dez-03
Summary
• Scenario: mobile user in intelligent “AR-ready”
building
• AR Toolkit used for
– High-precision tracking in small areas
– Coarse room-based wide area tracking ( “ID tags”)
• Software architecture for
– Dynamic context aware reconfiguration of AR Toolkit
marker detection
– Room-dependent choice of transition strategies
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
2
Outline
• Scenario: Intelligent Building
• System Components
– DWARF overview
– Configuration components
• Distributed Configuration
– Description of concepts
– Detailed example scenario
• Strategies for Using AR Toolkit as Wide Area
Tracker
• Conclusion & Future Work
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
3
Scenario: Intelligent Building
• User walks around in an
intelligent building,
wearing mobile client with
camera and AR Toolkit
• AR apps tied to specific
rooms, i.e. table-based AR
in Lab
• Mobile client has no a
priori knowledge of
environment
• Detection of current room
triggered by sets of Toolkit
markers
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
4
Scenario: Requirements
• No a priori knowledge about environment:
fi Distributed
architecture
fi Dynamic configuration
fi Split AR Toolkit functionality in components
• Trigger detection of room changes using AR
Toolkit markers
fi Allow
varying room-dependent strategies to increase
robustness
fi Provide flexible architecture to use additional
information sources (RF ID tags, ultra wideband
trackers, …)
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
5
Outline
• Scenario: Intelligent Building
• System Components
– DWARF overview
– Configuration components
• Distributed Configuration
– Description of concepts
– Detailed example scenario
• Strategies for Using AR Toolkit as Wide Area
Tracker
• Conclusion & Future Work
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
6
DWARF Overview
• Distributed Wearable Augmented Reality
Framework
• CORBA-based middleware dynamically connects
Services (DWARF components) based on
description of their Needs and Abilities
• Ability descriptions may be enhanced using
Attributes describing contextual information
• Many communication protocols
– CORBA method invocations
– CORBA notification service (events)
– Shared memory (for video transmission)
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
7
Components - Overview
Video Grabber: gets video image and puts it in a
shared memory segment
Marker Detection: takes video image out of shared
memory, detects AR Toolkit markers, sends
events describing position and ID of markers
Marker Loader: configures the Marker Detection
using remote method invocation with marker
descriptions suitable for the current context
Coarse Tracker: listens for events from Marker
Detection and changes the mobile client’s context
according to the new room
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
8
Components - Connections
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
9
Outline
• Scenario: Intelligent Building
• System Components
– DWARF overview
– Configuration components
• Distributed Configuration
– Description of concepts
– Detailed example scenario
• Strategies for Using AR Toolkit as Wide Area
Tracker
• Conclusion & Future Work
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
10
Distributed Configuration
• Basic Idea: distribute information processing
– Video Grabber and Marker Detection on user’s mobile
client
– Marker Loaders and Coarse Trackers in rooms
• Environmental services run in multiple instances,
one for each room
– Room-dependent evaluation of Marker Detection
events gets possible
– Available more accurate tracking technology (i.e. ultra
wideband tracker) can be integrated seamlessly
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
11
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
CoarseTracker
Load Room 1
Context Markers
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
Room 2 AR
App
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
MarkerDetection
Send marker events
Room 1
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
MarkerDetection
Send marker events
Detect Room
Change
Room 1
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
Change Client’s
Context to
Room=Room2
VideoGrabber
MarkerDetection
Send marker events
Room 1
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
MarkerDetection
Send marker events
Room 1
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
CoarseTracker
Unload Room 1
Context Markers
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
Room 2 AR
App
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
CoarseTracker
Load Room 2
Context + App
Markers
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
Room 2 AR
App
MarkerLoader
16-Dez-03
12
Distributed Configuration - Example
Mobile Client
VideoGrabber
Room 1
MarkerDetection
Room 2
CoarseTracker
Send marker events
CoarseTracker
Room 2 AR
App
MarkerLoader
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
MarkerLoader
16-Dez-03
12
Outline
• Scenario: Intelligent Building
• System Components
– DWARF overview
– Configuration components
• Distributed Configuration
– Description of concepts
– Detailed example scenario
• Strategies for Using AR Toolkit as Wide Area
Tracker
• Conclusion & Future Work
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
13
Strategies for Marker-Based Wide Area Tracking
Factors influencing the
choice of strategies:
Basic strategies:
•
•
•
•
•
• Minimum number of
frames
• Weighting number of
frames by confidence
value
• Detect multiple markers
simultaneously
• Detect sequence of
markers
Light conditions
Movement of user
Marker position
Camera
Room properties
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
14
Outline
• Scenario: Intelligent Building
• System Components
– DWARF overview
– Configuration components
• Distributed Configuration
– Description of concepts
– Detailed example scenario
• Strategies for Using AR Toolkit as Wide Area
Tracker
• Conclusion & Future Work
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
15
Conclusion & Future Work
• Prototypical implementation:
– Proof of concept for configuration architecture
– Demonstration within ARCHIE project
• Future Work:
–
–
–
–
–
Evaluate robustness of different strategies
Describe criteria for suitability of different strategies
Design combination strategies
Incorporate additional trackers/ information sources
Algorithms for detecting and reverting wrongly triggered
context changes
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
16
Thank you!
See our website:
http://www.augmentedreality.de
Or send E-Mail:
dwarf@augmentedreality.de
Configuration Strategies of an AR Toolkit-based
Wide Area Tracker
Martin Wagner, Felix Loew
16-Dez-03
17
Download PDF