Research
MOTAR
Teaching a drone to intercept a moving target — with nothing but its own sensors.
Why
Ordinary navigation goes to a place that stays put. Interception does not. The goal is a future position of something that is still moving, and every obstacle that hides the target also breaks the path that would have reached it.
The drone has to keep going on what it last saw, while the safe route and the intercepting route keep drifting apart.
How
A camera and LiDAR give it a view of the scene — no external tracking, no prebuilt map. A learned policy turns that view into the next maneuver, trained in simulation. A flight controller keeps each command inside what the aircraft can actually do.
Now
Simulation is running, and the corrected environment is verified. Physical target routing did not pass its gate. Hardware work has not started, and there have been no real-world flights.