Autonomous Navigation and Landing in

Urban Environments

Video Demonstration

Integrated stereo vision and Lidar sensors to accurately identify landing sites and navigate complex urban environments to perform dynamic landing maneuvers.

State Estimation & Self-Localization

Localization and tracking of landing-site is completed through sensor fusion of stereo camera and lidar data via an Unscented Kalman Filter (UKF)

Trajectory Generation and Tracking

System generates a trajectory avoiding nearby objects and within vicinity of targeted landing surface

Execution of Landing Maneuver

Robot executes DeepRL trained policy based on velocity and onboard state estimation