TL;DR
The MAGELLAN project tackles the complexity of creating multi-participant, mixed-reality location-based experiences by developing an integrated platform. This platform simplifies authoring through visual tools and natural interfaces, while enhancing the user experience with advanced AR and indoor localization techniques.
What Problem Does This Paper Address?
Creating engaging location-based experiences (LBEs) that blend the real and digital worlds for multiple users is currently a significant technical hurdle. These experiences, often involving mixed realities, demand sophisticated technologies and complex concepts, making them inaccessible to most creators. The MAGELLAN project aims to democratize the creation and enjoyment of these LBEs by providing a unified, user-friendly platform.
Methods
MAGELLAN employs a multi-faceted approach:
- Visual Authoring System: Built upon existing visual storytelling tools (INSCAPE), it's extended to support multi-participant AR experiences. This system allows non-programmers to author experiences by modeling them as a graph of activities, controlled by participant location and interactions.
- Natural User Interface: This includes tangible objects that can be used with interactive surfaces for intuitive control of the authoring tool, with the added ability to create custom tangible objects via 3D printing.
- Seamless Hybrid Positioning System: This system combines GPS, WLAN, and FMCW radar for accurate indoor and outdoor localization, aiming for optimal balance between accuracy and cost.
- Panorama-Based AR Tracking: A novel machine learning-based feature point detector is used, outperforming traditional methods like SIFT in robustness to illumination changes and repeatability for reliable AR tracking in diverse outdoor conditions.
- Context-Aware Augmented Reality: Research into how context influences AR utility is being conducted, with an approach for augmenting public displays without extensive infrastructure, using real-time streaming protocols.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Democratizing LBE Creation: The core idea is to abstract away the underlying technical complexities, enabling a wider range of creators to build rich, interactive experiences.
- Bridging Real and Digital: The project emphasizes merging the physical and digital worlds, not just overlaying digital information, but creating truly mixed realities.
- Intuitive Authoring: The use of visual tools and tangible interfaces is key to making the authoring process accessible and enjoyable.
- Robust AR in the Wild: Developing AR tracking that works reliably in uncontrolled, real-world environments (varying light, seasons) is crucial for practical applications.
- Public Displays as AR Entry Points: Leveraging existing public displays offers a scalable and accessible way to distribute AR experiences.
Key Results
The paper presents the first results of the MAGELLAN project, highlighting:
- A proof-of-concept system that has demonstrated the ability for non-programmers to create and publish complex VR/AR experiences in a matter of hours, a task that would typically require multiple technical experts.
- A novel feature point detector for AR tracking that shows improved robustness and repeatability compared to SIFT, enabling more reliable AR experiences in challenging outdoor conditions.
- The potential for 1 cm position accuracy and a 50% reduction in power consumption for mobile units using their developed localization techniques.
Limitations and Future Work
- Proof-of-Concept: The system is currently a proof-of-concept, with a beta version planned for release.
- Future Work: The authors intend to explore combining AR with alternative interfaces for seamless interaction, both on and away from physical artifacts. Further development of the positioning system to incorporate additional standards is also planned.
Summary updated