03/04/2026
🚀Unlock Next-Level State Estimation with VOXL Open-VINS Server!🚀
We are excited to spotlight the voxl-open-vins-server, our powerful multi-camera visual-inertial odometry (VIO) service hardware-optimized for the VOXL 2 platform!
Built around an advanced MSCKF-based estimator derived from OpenVINS, this service seamlessly fuses data from an IMU and up to 3 tracking cameras. It processes this data to deliver real-time pose and velocity estimates directly to your autopilot, ensuring rock-solid odometry for your autonomous flights.
With our recent VOXL SDK 1.6.3 release, the voxl-open-vins-server has been upgraded to version 0.6.0, bringing even better multi-camera VIO performance to your drone.
Here is why developers love it:
🔹Multi-Camera Fusion: Maximize your tracking reliability by feeding the estimator with synchronized frames from up to 3 mono or stereo tracking cameras simultaneously.
🔹Smart Occlusion Handling: Have a downward-facing tracking camera? The server can be configured to ignore features from occluded cameras while the drone is on the ground, automatically activating them once it reaches your specified takeoff altitude.
🔹Robust Auto-Recovery: Built-in health monitoring runs at 30 Hz to continually check system connectivity and tracking quality. If divergence or a low feature count is detected, the system features configurable auto-reset and fallback modes to keep your drone safe.
🔹MPA Architecture: The server utilizes Modal Pipe Architecture (MPA) to efficiently decouple sensor producers from VIO consumers, making your multi-camera VIO pipeline highly modular and optimized.
Upgrade to VOXL SDK 1.6.3 today and experience the most robust visual-inertial odometry available for the VOXL 2!
Find out more about SDK 1.6.3 here:
ModalAI technical documentation for VOXL and VOXL 2 Companion Computers for PX4 and ArduPilot Obstacle Avoidance and GPS-denied navigation, assembled in the USA
