GNSS Denied Navigation

Point and Track locks onto a designated point, feature, or dynamic object in the live camera feed and continuously maintains line-of-sight tracking during flight maneuvers. Correlation algorithms, optical centroid tracking, and deep-learning object detection maintain target lock despite rapid viewpoint shifts, scale changes, background clutter, and partial occlusions.

Tracking deviations feed directly into gimbal stabilization servos and autopilot flight guidance loops. This enables autonomous line-of-sight following, orbit patterns, and persistent surveillance without depending on the target's GPS coordinates or external communication links.

Autonomous Target Engagement closes flight guidance and control loops directly around real-time visual tracking errors during the terminal phase of a mission. The autopilot executes proportional navigation (PN) and line-of-sight rate guidance laws derived from camera measurements, commanding aerodynamic surfaces and thrust vectors to minimize the miss distance.

This capability is vital for loitering munitions, precision kinetic interception, and automated landing on moving platforms (such as ships or land vehicles), ensuring successful mission completion even when electronic warfare jammers completely deny satellite navigation and ground communications.

Inertial Navigation Systems (INS) provide autonomous dead-reckoning state estimation without external radio signals. Using a tactical-grade MEMS Inertial Measurement Unit (IMU) the system measures specific forces and angular rates in real time. The navigation processor integrates angular velocity using singularity-free quaternions for 3D attitude tracking, transforms body accelerations into the local NED frame, and compensates for gravity and Coriolis effects.

An Extended Kalman Filter (EKF) continuously estimates sensor biases and fuses inertial propagation with secondary aids: barometric altimetry and magnetometers stabilize altitude and heading, while optical flow or visual-inertial odometry constrains velocity drift. For prolonged operations, absolute position drift is bounded using map-matching or terrain relative navigation (TRN).

Radar Terrain Relative Navigation (R-TRN) utilizes active radio-frequency sensors—such as radar altimeters or imaging radar—to sample terrain elevation contours and match them against onboard Digital Elevation Models (DEM / DTED). Unlike optical cameras that can be blinded by fog, clouds, smoke, or total darkness, radar wavelengths penetrate adverse weather conditions and obscurants.

Terrain contour correlation algorithms match the radar profile to the pre-stored elevation grid, deriving absolute horizontal coordinates to bound inertial drift and provide dependable, all-weather navigation in zero-visibility environments.

Optical Landing Spot Detection enables UAVs to land autonomously in GNSS-denied environments by using downward-facing cameras to identify specific visual or thermal ground markers. This creates a reliable visual reference point regardless of lighting conditions or GPS jamming.

Once locked on, the system calculates the drone's 6-DoF pose—its exact 3D position and orientation—in real time. These rapid, continuous calculations allow the UAV to dynamically adjust its descent and safely synchronize with moving vehicles or pitching ship decks.

Harnessing the power of Optical Flow, our system analyzes ground imagery in real time to track the displacement of visual features. This provides a completely passive and highly accurate estimation of the aircraft’s movement (Visual Odometry).

Rather than relying on external signals, the autopilot seamlessly fuses this visual data with its internal inertial and altitude sensors. This advanced sensor fusion instantly translates pixel motion into true metric ground velocity.

The result is a robust, self-contained navigation solution. It empowers the aircraft to eliminate horizontal drift, maintain precision hovering, and execute agile maneuvers in strictly GPS-denied environments—all while remaining completely stealthy and emission-free.

Visual Simultaneous Localization and Mapping (vSLAM) enables autonomous platforms to navigate unmapped, GPS-denied environments by constructing a 3D map of the surroundings in real time while tracking the aircraft's own location within that map. A tracking front-end estimates short-term trajectory across sequential frames, while an optimization back-end applies Bundle Adjustment to keep the map and camera poses structurally consistent.

A crucial advantage is Loop Closure Detection: when the aircraft revisits previously mapped areas, visual fingerprinting algorithms identify the match and execute pose-graph optimization. This distributes and cancels accumulated dead-reckoning drift across the entire mission history.

Visual Map-Matching (VMM) provides passive, drift-free absolute global coordinates by matching live aerial camera imagery against pre-loaded, geo-referenced satellite or drone orthomosaics stored in the onboard database. Live frames undergo real-time perspective rectification based on the aircraft's estimated roll, pitch, and altitude.

Robust invariant feature descriptors extract permanent structural landmarks—such as road intersections, riverbanks, coastlines, and runway outlines—that remain identifiable across varying lighting and weather conditions. Solving the image-to-map transformation matrix yields exact latitude and longitude fixes, bounding inertial drift across long-range flights with zero radio emissions.

LEO Positioning leverages satellite constellations orbiting in Low Earth Orbit (500 to 1,500 km altitude) to provide jam-resistant positioning, navigation, and timing (PNT) as an alternative or backup to traditional GNSS. Because LEO satellites operate much closer to Earth than MEO constellations (~20,000 km), their signal power at the aircraft antenna is 20 to 30 dB higher, making them substantially harder to jam.

Additionally, LEO satellites travel at high orbital velocities with rapid geometry changes and pronounced Doppler shifts. The navigation receiver measures Doppler frequency variations and ranging signals to compute velocity vectors and position fixes independently of legacy GPS or Galileo systems.

• Low Earth Orbit PNT
• Alternative Satellite Navigation

4G/LTE PNT provides an independent, terrestrial alternative positioning, navigation, and timing source that operates without any reliance on satellite constellations. Utilizing an onboard cellular transceiver connected to the flight control architecture, the system listens to ambient radio frequency emissions broadcast by commercial multi-tower telecommunication networks, identifying nearby base stations through unique cell IDs.

By measuring Time-Difference-of-Arrival (TDOA), Received Signal Strength (RSSI), and antenna phase Angle-of-Arrival (AOA) relative to known terrestrial tower coordinates, the navigation filter executes real-time hyperbolic triangulation. This yields a reliable horizontal position fix to bound inertial drift and maintain safe trajectory control throughout low-altitude corridors and urban canyons.

• Cellular Network Positioning
• Terrestrial Radio Ranging
• Signals of Opportunity (SoOP) PNT