Learned FBL
Flatness-Preserving Residual Learning for Real-Time Tight Quadrotor Formation Flight
By leveraging differential flatness, we propose Learned FBL, a framework that enables a team of quadrotors to fly in tight formations while using an order of magnitude less computational cost than the state-of-the-art KNODE-DW MPC.
Publication Link: IROS 2026
Publication PDF: PDF
Demo video at: Demo
Time-lapse of two Crazyflie quadrotors using our Learned FBL framework merging from a 1.1 m stacked formation into a tight 0.1 m separation to fly through a 0.4 m high window.