Learning Agile Quadrotor Flight in the Real World (RSS 2026) @ailabRPG
Learning Agile Quadrotor Flight in the Real World (RSS 2026)  @ailabRPG
Uploaded July 2026 | Updated November 2026, 10 hours ago
Learning-based controllers have achieved impressive performance in agile quadrotor flight but typically rely on large scale simulation training, necessitating accurate system identification for effective sim-to-real transfer. However, even with precise modeling, fixed policies remain susceptible to out-of-distribution scenarios, ranging from external aerodynamic disturbances to internal hardware degradation. To ensure safety under these evolving uncertainties, such controllers are forced to operate with conservative safety margins, inherently constraining their agility outside of controlled settings. While online adaptation offers a potential remedy, safely pushing the platform toward its agility envelope remains a critical bottleneck due to data scarcity and safety risks. To bridge this gap, we propose a self-adaptive framework that eliminates the need for precise system identification or offline sim-to-real transfer. We introduce Adaptive Temporal Scaling (ATS) to actively push the platform toward the agility ceiling exposed by its control interface, and employ online residual learning to augment a simple nominal model. Based on the learned hybrid model, we further propose Real World Anchored Short-Horizon Backpropagation Through Time (RASH-BPTT) to achieve efficient and robust in-flight policy updates. Extensive experiments demonstrate that our quadrotor reliably executes agile maneuvers near the saturation limit of the CTBR command interface. The system evolves a conservative base policy from a peak speed of 2.0 m/s to 7.3 m /s within approximately 100 s of flight time. These findings underscore that real-world adaptation serves not merely to compensate for modeling errors, but as a practical mechanism for sustained
performance improvement in aggressive flight regimes.

Reference:
R. Ren, Z. Zhu, J. Xing, D. Scaramuzza,
"Learning Agile Quadrotor Flight in the Real World",
Robotics: Science and Systems (RSS), 2026, Sydney, Australia
PDF: https://rpg.ifi.uzh.ch/docs/RSS26_Ren.pdf
Project Page: https://rpg.ifi.uzh.ch/lafr/
Code: github.com/uzh-rpg/LAFR

For more info about our research on:
Agile Drone Flight: http://rpg.ifi.uzh.ch/aggressive_flight.html
Machine Learning: http://rpg.ifi.uzh.ch/research_learning.html

Affiliation:
Yunfan Ren, Zhiyuan Zhu, Jiaxu Xing, and Davide Scaramuzza are with the Robotics and Perception Group, University of Zurich, Switzerland: https://rpg.ifi.uzh.ch/
Learning Agile Quadrotor Flight in the Real World (RSS 2026)Learning Visuomotor Flight Through Dynamic Narrow Gap via Differentiable Simulation (CoRL 2026)Time-Optimal Online Replanning for Agile Quadrotor Flight (RAL 2022)Data-Efficient Decentralized Visual SLAM (ICRA18 Video Teaser)A Benchmark Comparison of Learned Control Policies for Agile Quadrotor Flight (ICRA 2022)
UZH Robotics and Perception Group |

Learning Agile Quadrotor Flight in the Real World (RSS 2026)

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER