TC1: Safe Perception, Coordination, Planning, and Navigation

The full-scale incorporation of low-altitude UAVs and UAM systems introduces unexplored safety risks, and necessitates high-reliability sensing, coordination, planning, and navigation approaches to cope with dynamic and unpredictable environments. TC1 addresses the amalgamation of various algorithms (and software components) that correspond to disparate technological challenges, operate at different time-scales, and integrate dissimilar types of data into a unified, holistic format compatible with the existing and near-future embedded flight control systems for UAV and UAM systems.

Urgent Need for Safe Perception, Coordination, Planning, and Navigation

TC1 will develop computationally effective and scalable sensing, perception, coordination, planning, and learning-based control techniques, to provide a safe and scalable UAM network. This will lead to future UAM networks that reliably and adaptively provide mobility services for spatio-temporally distributed demands over a congested and potentially GPS-degraded airspace cluttered by moving and unknown objects. Near-future UAM networks are expected to involve both UAV-to-Ground station (U2G) and UAV-to-UAV (U2U) communication capabilities.

To achieve the TC1 overarching goal, we will complete the following objectives

  • (Objective I) Sensing, Perception, and Semantic Scene Understanding
  • (Objective II) On-the-Fly Conflict-Free Tasking and Coordination
  • (Objective III) Compositional Learning for Safe Motion Planning Control and Navigation