Learning Agile Intruder Interception using Differentiable Quadrotor Dynamics
Enables agile quadrotor interception with passive monocular cameras, addressing a practical sensing limitation in real-world drone pursuit tasks.
The paper proposes a control policy for quadrotor interception of intruders using only 3D direction vectors, learned via differentiable dynamics, achieving 30% better performance than baselines at speeds up to 10 m/s.
This paper presents a methodology for learning a control policy to intercept an intruder using the 3D direction unit vector to the intruder and the interceptor state. Prior deep reinforcement learning approaches assume either relative position or distance to the intruder is available, but this information is not readily accessible in real-world applications that employ passive, monocular camera sensors. Instead, we propose a solution that leverages an analytical policy gradient method using differentiable quadrotor dynamics to learn agile interception at speeds up to 10 m/s. The proposed approach outperforms baseline methods that utilize simplified point mass dynamics by an average of 30%.