Spatial applications, i.e., applications that tie digital information with the physical world, have improved many of our daily activities, such as navigation and ride-sharing. This class of applications also holds significant promise of enabling new industries such as augmented reality and robotics. The development of these applications is enabled by spatial intelligence systems, or systems that can provide physically-grounded services such as routing, spatial search, and localization. Today, mapping platforms provided by organizations like Google and Apple serve as spatial intelligence systems. These maps are centralized and primarily cover outdoor spaces. We envision that future spatial applications, such as persistent world-scale augmented reality and robotics, would require detailed and precise spatial intelligence across indoor and outdoor spaces. The scale of cartography efforts required to survey indoor spaces and their privacy needs inhibit existing centralized maps from incorporating such spaces into their platform.

This thesis proposes the design and implementation of a federated spatial intelligence platform. It first introduces OpenFLAME (Open Federated Localization and Mapping Engine), a federated mapping infrastructure where independent parties can manage and serve their own maps of physical regions. This unlocks scalability of map management, isolation, and privacy of maps. This is implemented on top of the existing Domain Name System (DNS), which enables us to leverage its existing infrastructure. Secondly, it explores the development of federated spatial services with a focus on the localization service. Finally, we explore ways to leverage latest advancements in AI to make it easier to build indoor spatial services by automatically extracting structure from raw spatial data (e.g., 3D scans and video walkthroughs). We will also go beyond traditional services to build a spatial reasoning service to answer free-form questions about the space.

We are increasingly seeing evidence of ubiquitous deployment of spatial devices like robots and augmented reality glasses in the near future. Such devices need an infrastructure to contextualize themselves in their current environments. This thesis provides the foundations for such an infrastructure.

98 pages

Thesis Committee:
Srinivasan Seshan (Co-chair)
Anthony Rowe (Co-chair)
Justine Sherry
Hari Balakrishnan (Massachusetts Institute of Technology)

Jignesh Patel, Interim Head, Computer Science Department
Martial Hebert, Dean, School of Computer Science

Return to: SCS Technical Report Collection
School of Computer Science

This page maintained by reports@cs.cmu.edu