Learning Robust Agile Flight Control with Stability Guarantees

2025-10-14

AI Summary

Achieving high-precision trajectory tracking for agile quadrotors under actuator saturation, external disturbances, and strict real-time constraints remains challenging. Method: This paper proposes a neural-augmented feedback controller with rigorous stability guarantees, integrating a nonlinear feedback structure, deep reinforcement learning, and Lyapunov stability theory—deployable directly on physical platforms without Sim-to-Real fine-tuning. Contribution/Results: The controller explicitly incorporates actuator limits while ensuring closed-loop stability; a lightweight neural network compensates online for modeling errors and strong disturbances. Operating at ≥500 Hz with low computational overhead, it achieves accurate tracking of highly aggressive trajectories. Experiments demonstrate superior robustness and real-time performance under wind gusts and other severe disturbances, outperforming conventional nonlinear and purely learning-based controllers.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Multiagent Systems: Multiagent Learning
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

In the evolving landscape of high-speed agile quadrotor flight, achieving precise trajectory tracking at the platform's operational limits is paramount. Controllers must handle actuator constraints, exhibit robustness to disturbances, and remain computationally efficient for safety-critical applications. In this work, we present a novel neural-augmented feedback controller for agile flight control. The controller addresses individual limitations of existing state-of-the-art control paradigms and unifies their strengths. We demonstrate the controller's capabilities, including the accurate tracking of highly aggressive trajectories that surpass the feasibility of the actuators. Notably, the controller provides universal stability guarantees, enhancing its robustness and tracking performance even in exceedingly disturbance-prone settings. Its nonlinear feedback structure is highly efficient enabling fast computation at high update rates. Moreover, the learning process in simulation is both fast and stable, and the controller's inherent robustness allows direct deployment to real-world platforms without the need for training augmentations or fine-tuning.

Problem

• Achieving precise trajectory tracking at operational limits
• Handling actuator constraints and robustness to disturbances
• Providing universal stability guarantees for agile flight

Innovation

• Neural-augmented feedback controller for agile flight
• Universal stability guarantees under extreme disturbances
• Direct simulation-to-reality deployment without fine-tuning