Autonomous Cloud Avoidance

Passive Camera Based Cloud Detection and Avoidance

This technology is the first practical solution for the problem of maintaining UAS in Visual Meteorological Conditions (VMC) when on-board radar is not available. Our software processes data from on-board GPS, IMU, passive imaging sensors and extracts 3D cloud formation information. Autonomous avoidance of clouds maintains UAS operations in VMC, thereby ensuring due-regard for the safety of other air traffic. This technology fulfills a critical need for a passive-only Sense-and-Avoid solution for UAS.


Technology description

  • Detection and extraction of 3D cloud formation data using passive Electro-Optical/Infrared imaging sensors, on-board GPS and IMU sensors.
  • Developed a Jointly Optimal Cloud and Collision Avoidance (JOCCA) framework in close collaboration with Northrop Grumman Corporation.
  • JOCCA performs UAS waypoint modification while ensuring simultaneous avoidance of clouds and air traffic.
  • Alternate waypoints are generated without violating the kinodynamic constraints of the UAS.
  • Follows FAA cloud avoidance regulations, including limits on vertical and lateral clearance.
  • Tested and shown to perform well in a high-fidelity simulation environment with synthetic and recorded flight data.
  • Current implementation runs on a small form factor PC104 single board computer at 1 Hz.