Project 'ATTENTION': Advanced TargeT dEtectioN, recogniTion and Identification based on multispectral inertial Odometry for Navigation

Global Navigation Satellite System (GNSS) is a vital technology for Unmanned Airial Vehicles (UAVs), providing accurate positioning and enabling autonomous navigation. However, in hostile environments, GNSS signals can be jammed or unavailable, posing a significant security threat and making reliable operation in GNSS-denied areas a critical challenge. Though equipping UAVs with active sensors such as LiDAR and radar enhances navigational autonomy and perception, they compromise mission stealth and covert operations as their emitted signals can be detected by enemy surveillance systems. Passive sensing on the other hand provides a stealthier alternative, relying on non-emitting sensors to capture environmental data. However, monomodal passive systems face limitations due to their narrow spectral range, which affects their ability to detect, recognize, and identify targets, especially under challenging conditions like smoke, adverse weather, poor lighting, or camouflage. These systems are also prone to false positives and negatives, reducing their reliability in complex scenarios. While Inertial Navigation Systems (INS) have demonstrated effectiveness in GNSS-denied environments, such as underwater, their utility in UAVs is hindered by cumulative drift errors over time. ATTENTION seeks to overcome all these limitations by advancing beyond current solutions like Skynode, an all-in-one autopilot software stack flight controller with real-time kinetic GNSS, IMU, telemetry and an onboard neural-processing unit for camera-based vision.