Event
Reinforcement-learning fixed-wing interceptor completes real flight validation
1 reports · 1 sources
A team at Delft University of Technology in the Netherlands proposed a fixed-wing interceptor drone controlled by a neural network: the network maps filtered monocular visual observations directly to actuator commands, and the policy is trained with reinforcement learning and proximal policy optimization (PPO) in a custom Python environment. In simulation, the policy outperformed two classical guidance controllers developed concurrently by the team under most conditions; the policy was then deployed to a fixed-wing platform and successfully intercepted a maneuvering fixed-wing target in real flight. This work provides validation for autonomous control of fixed-wing interceptor drones.
Report timeline
- Reinforcement Learning-Based Autonomous Fixed-Wing Drone Interception
A team at Delft University of Technology in the Netherlands proposed a fixed-wing interceptor controlled by a neural network that maps filtered monocular vision observations directly to actuator commands; the policy was trained in a custom Python environment using reinforcement learning and proximal policy optimization (PPO). In simulation, the policy outperformed two classical guidance controllers under most conditions, and it has successfully intercepted maneuvering fixed-wing targets in real flight. This work provides validation for the autonomous control of fixed-wing interception drones.
arXiv cs.RO ·