Robust image-based control for aerial interaction under external disturbances

Area of Science:

• Robotics
• Control Systems
• Computer Vision

Background:

• Aerial manipulators (AMs) face challenges in infrastructure inspection due to coupling, uncertainties, and disturbances.
• Robust control is essential for precise contact inspection and aerial physical interaction (APhI).

Purpose of the Study:

• To develop a robust image-based control framework for AMs performing contact inspection.
• To enhance force-tracking accuracy and robustness against disturbances and constraints.

Main Methods:

• Derived a decoupled second-order image dynamics model using virtual-camera image moments.
• Developed a disturbance observer-based nonlinear model predictive controller (NMPC) with online disturbance estimation and constraint enforcement.
• Designed an image-based variable-stiffness impedance controller for force-tracking and a geometric attitude controller to prevent Euler angle singularities.

Main Results:

• Achieved significant reduction in contact force Root Mean Square Error (RMSE): 31.6% in simulation and 58.3% in experiments.
• Demonstrated improved tracking accuracy and robustness in aerial physical interaction (APhI) tasks.
• Verified force-tracking accuracy and reproducibility across multiple trials.

Conclusions:

• The proposed image-based control framework enhances aerial manipulator performance in contact inspection.
• The system offers superior robustness and force-tracking capabilities compared to existing methods.
• Validated the framework's effectiveness for reliable aerial physical interaction (APhI) applications.