One fusion hub, every sensor, a stock flight controller
Actively being built. The design is scoped and firm; it is not usable yet.
How it works
Sensors in, one estimate out
Every sensor publishes through its own driver. A shared fusion layer combines them and speaks standard MAVLink to the flight controller.
Layers
Three layers on the companion computer
Each layer is open and replaceable. Sensors are drivers, fusion is a shared estimate, and the mapping keeps the flight controller stock.
The organizing idea
Reliability while moving
Sensors that stay accurate in flight carry more weight. Sensors that get noisy with speed or texture are trusted less as the drone moves.
Motion-agnostic
Steady in flight, so they carry high weight in the estimate.
- ToF rangefinder
- RTK GNSS
- Solid-state lidar
- mmWave radar
Scene-sensitive
Reliable when slow and textured, so their weight drops with speed and yaw.
- Optical flow
- Stereo depth
- Mechanical 2D lidar
Mapping
Sensor roles become standard messages
The fusion layer never forks firmware. It emits the MAVLink messages a stock flight controller already reads.
- Rangefinder
- DISTANCE_SENSOR
- Obstacle field
- OBSTACLE_DISTANCE
- 3D obstacle field
- OBSTACLE_DISTANCE_3D
- Visual position
- VISION_POSITION_ESTIMATE
- Odometry
- ODOMETRY
Related
Explore the stack
One ecosystem, any sensor
Sensor drivers, fusion, and MAVLink mapping ship as open-source. Follow the design as it comes together.
Read the docs