UAVOS and Mira Aerospace have completed a flight test of NAVAI, a vision-based navigation module designed to provide continuous positioning for Unmanned Aircraft Systems (UAS) in GNSS-denied conditions.
Integrated into the ApusNeo18 High-Altitude Pseudo-Satellite (HAPS), NAVAI maintained precise navigation at an altitude of 16,000 meters, approximately 52,500 feet.
Unmanned aircraft typically rely heavily on GNSS for positioning and navigation. When GNSS availability or accuracy is affected by jamming, spoofing or natural interference, NAVAI provides an alternative capability by processing video from onboard cameras together with aircraft telemetry.
Inertial Navigation Systems (INS) can also provide backup navigation, but accumulated drift can cause position errors to increase over time. NAVAI is designed to address this limitation by combining visual odometry with terrain-referenced navigation.
The module compares real-time camera imagery with pre-loaded reference imagery to generate a bounded, non-drifting absolute position estimate when suitable terrain is visible. At high altitudes, where terrain features may be sparse, NAVAI can obtain a position fix from a single frame.
Onboard edge computing and neural networks enable the system to process visual data in real time. Neural network feature matching identifies distinctive ground features and compares them with an onboard database of terrain maps to determine the aircraft’s location.
AI-based noise filtering is used to compensate for visual disturbances including cloud cover, haze and low-light conditions, as well as seasonal changes such as snow cover and vegetation growth.
A dedicated Ground Control Station interface also enables operators to view a live overlay of the aircraft’s visual matches directly on the mission map.
Aliaksei Stratsilatau, Founder and CEO of UAVOS, commented, “Testing NAVAI on the ApusNeo platform reflects our commitment to making UAS operations truly autonomous and resilient. By giving the aircraft the ability to ‘see’ and understand its surroundings, we are eliminating a critical dependency on satellite navigation, which has historically limited drone operations in contested and GNSS-denied environments.”
