Inertial Labs, a VIAVI Solutions company, has released a case study detailing the integration of its Visual Inertial Navigation System (VINS) with Vantor’s Raptor Guide™ Vision-Based Positioning Software. Read more >>
The study examines how combining inertial navigation with vision-based processing enables reliable positioning in environments where GNSS signals are unavailable or degraded, supporting continued operation across demanding mission scenarios.
As autonomous systems are deployed in increasingly complex and signal-restricted environments, reliance on GNSS alone presents operational limitations. The case study outlines how VINS addresses this challenge by fusing inertial measurement data with visual inputs, enabling continuous position updates and reducing dependency on external infrastructure. This combined approach allows the system to maintain localization during GNSS outages through the integration of visual inertial odometry and vision-based map-matching techniques.
Flight testing conducted using a fixed-wing aircraft evaluated system performance across visually challenging terrain, including mountainous regions and mixed land features. During GNSS-denied segments, the system maintained bounded positional error with controlled drift, despite variations in visual conditions. Results showed a mean position error of 12.20 meters, with CEP50 and CEP95 values of 9.31 meters and 25.81 meters respectively, demonstrating stable and repeatable performance across multiple flight profiles.
The case study highlights the role of tightly coupled sensor fusion, where vision-based positioning and visual inertial odometry provide aiding data to the inertial navigation system to sustain accuracy when satellite signals are unavailable. The results demonstrate how multi-sensor integration supports resilient navigation for applications such as UAVs, robotics, and defense platforms operating in contested or constrained environments.
Download the full technical report to explore the complete analysis and flight test results.
