MIDIH Research Studio-AWEAR experiment
- 1. AWEAR – A Wearable Expert
Augmented Reality System
Speaker: Christian Thomay – RSA FG
- 3. AWEAR
AWEAR – A Wearable Expert Augmented Reality System
The AWEAR platform enables the creation of 3D maps of complex
industrial facilities with low-cost mobile sensors.
AWEAR platform is able to accurately
localize the worker (position and orientation)
without the need of further expensive sensing
infrastructure and provide assistance in the
form of augmented navigation guidance.
- 5. Point Cloud Generation
• Generate 3D point cloud of industrial environment
• Use smart phones with depth cameras
• Point cloud generated in RTAB-Map software
- 6. RGB-Based Localization
• Localize user live in 3D environment
• Utilize only camera RGB image in ACG-Localizer software
• Localizer builds vocabulary of visual words, finds position in 3D space based on RANSAC word search
• Successful localization about every 3-4 seconds
Illustration of localization based on 2D image features [from Sattler et al]
- 7. Location Tracking
• Since localization is only successful every 3-4
seconds, need to track user position in
between
• Use Visual Inertial Odometry
• Tracks motion of the smart phone using RGB
camera image and data from inertial sensors
• Integrated Google ARCore for this
- 8. User Guidance
• User is localized and guided
to location in facility using AR
markers
• AR markers implemented in
Unity engine
• App with AR markers
deployed to smart phone
• Markers overlayed over
phone camera image
- 12. Conclusions and Outlook
AWEAR platform successfully shown ability to
• Generate a 3D model of industrial space
• Localize user in mapped-out space
• Guide user to location using AR interface
Future exploitation plan in Augmented Maintenance Support:
I. Pathfinding / Navigation (current solution)
II. Task Execution Support: highlight parts, indicate work steps etc.
III. Automated Documentation: digital documentation of performed work steps using AR
assistance systems