Learning Real-World Acrobatic Flight from Human Preferences
By: Colin Merk , Ismail Geles , Jiaxu Xing and more
Potential Business Impact:
Teaches drones to fly fancy tricks by watching.
Preference-based reinforcement learning (PbRL) enables agents to learn control policies without requiring manually designed reward functions, making it well-suited for tasks where objectives are difficult to formalize or inherently subjective. Acrobatic flight poses a particularly challenging problem due to its complex dynamics, rapid movements, and the importance of precise execution. In this work, we explore the use of PbRL for agile drone control, focusing on the execution of dynamic maneuvers such as powerloops. Building on Preference-based Proximal Policy Optimization (Preference PPO), we propose Reward Ensemble under Confidence (REC), an extension to the reward learning objective that improves preference modeling and learning stability. Our method achieves 88.4% of the shaped reward performance, compared to 55.2% with standard Preference PPO. We train policies in simulation and successfully transfer them to real-world drones, demonstrating multiple acrobatic maneuvers where human preferences emphasize stylistic qualities of motion. Furthermore, we demonstrate the applicability of our probabilistic reward model in a representative MuJoCo environment for continuous control. Finally, we highlight the limitations of manually designed rewards, observing only 60.7% agreement with human preferences. These results underscore the effectiveness of PbRL in capturing complex, human-centered objectives across both physical and simulated domains.
Similar Papers
Best Policy Learning from Trajectory Preference Feedback
Machine Learning (CS)
Teaches AI to learn better from people's choices.
FLoRA: Sample-Efficient Preference-based RL via Low-Rank Style Adaptation of Reward Functions
Robotics
Teaches robots new tricks without forgetting old ones.
PB$^2$: Preference Space Exploration via Population-Based Methods in Preference-Based Reinforcement Learning
Artificial Intelligence
Teaches computers to learn from choices better.