Dynamic SLAM for Real-Time Drone Navigation and Immersion
Abstract
In many military and civilian applications that concern autonomous or semi-autonomous robots,the robots may share the same physical" space and further interact extensively with their humanusers. One of the key enabling technologies is simultaneous localization an"d mapping (SLAM). Inan existing ONR MURI program, our team has developed new theories and systems that addressthe traditional SLAM"" within the static environment. Yet, the Dynamic SLAM problem hasreceived much less attention, namely, the 3D scene is comprised of" a static environment and oneor more moving objects also within the environment. We propose to instrument a novel DynamicSLAM platform with the goal to achieve real-time dense 3D mapping and localization for rapidlymoving vehicles such as aerial drones. The reconstruction of the world geometry and the dynamicsof the moving objects within can become a part of new solutions to design new human-robotinteraction to effectively control the robot remotely through virtual/augmented reality and viahigh-level natural language and gesture interfaces. This abstract is publically releasable.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 01, 2017
- Source ID
- N000141712740
Entities
People
- Sosale Sastry
Organizations
- Office of Naval Research
- United States Navy
- University of California Regents