AirBender lets two drones transport bendable objects on the fly

AirBender tackles a fundamental challenge in aerial robotics: automated transport of deformable objects. Traditional controllers struggle with bendable payloads because they rely on precise elasticity models that are often unknown or hard to compute. In this work, Jiawei Xu and colleagues from the University of New Mexico and Lehigh University present an adaptive control strategy that lets two multirotor drones collaboratively carry a flexible object without needing those models. The system uses Lyapunov stability analysis to prove that the adaptive controller is asymptotically stable, meaning it reliably converges to desired trajectories even as the object deforms unpredictably. The drones adjust their forces and positions in real time based only on the object's motion, not its internal properties.

The method was validated through hardware experiments across multiple scenarios, including carrying a thin rod and a flexible sheet. The drones maintained stable flight and accurate trajectory tracking without prior knowledge of the object's stiffness or damping. This represents a significant step toward practical aerial manipulation of everyday flexible items. Potential applications include drone-based construction (moving hoses or cables), disaster response (carrying folded tarps), or warehouse logistics (transporting flexible packaging). The paper is published in IEEE Robotics and Automation Letters (DOI: 10.1109/LRA.2025.3536276) and available on arXiv (2605.07003).