RoVerFly: Robust and Versatile Implicit Hybrid Control of Quadrotor-Payload Systems
This work addresses the problem of adaptive and robust control for quadrotor-payload systems, which is crucial for applications like aerial transport and inspection, but it is incremental as it builds on existing learning-based methods for hybrid dynamics.
The paper tackles the challenge of designing robust controllers for quadrotors with flexible cable-suspended payloads, which involve nonlinear and hybrid dynamics, by presenting RoVerFly, a learning-based framework that uses a single reinforcement learning policy as an implicit hybrid controller. It achieves strong zero-shot generalization across varying payload settings, including no payload and changes in mass and cable length, without requiring re-tuning.
Designing robust controllers for precise trajectory tracking with quadrotors is challenging due to nonlinear dynamics and underactuation, and becomes harder with flexible cable-suspended payloads that add degrees of freedom and hybrid dynamics. Classical model-based methods offer stability guarantees but require extensive tuning and often fail to adapt when the configuration changes-when a payload is added or removed, or when its mass or cable length varies. We present RoVerFly, a unified learning-based control framework where a single reinforcement learning (RL) policy functions as an implicit hybrid controller, managing complex dynamics without explicit mode detection or controller switching. Trained with task and domain randomization, the controller is resilient to disturbances and varying dynamics. It achieves strong zero-shot generalization across payload settings-including no payload as well as varying mass and cable length-without re-tuning, while retaining the interpretability and structure of a feedback tracking controller. Code and supplementary materials are available at https://github.com/mintaeshkim/roverfly.