Hook-Based Aerial Payload Grasping from a Moving Platform

TL;DR

This work addresses autonomous payload grasping from a moving platform using a hook-based aerial manipulator. It integrates digital twin-based payload motion prediction with a computationally efficient trajectory optimization that embeds complementarity constraints to automatically select the optimal grasp time $T_g$ within a window $0<\underline{T}_g\le T_g\le \overline{T}_g< T_f$, solved via SQP-RTI, and an IQC-based robustness analysis to certify performance under uncertainties, all tracked by an LTV-LQR controller. The approach is validated through MuJoCo simulations and real flight experiments on a custom aerial manipulator platform, achieving reliable grasping despite dynamic terrain and disturbances; the robustness analysis yields a radius $\rho^*$ that bounds allowable hook deviations, ensuring grasp success when $r_{hook}>\rho^*$. Overall, the paper demonstrates a practical, real-time capable framework for hook-based aerial payload grasping from moving platforms with formal guarantees and demonstrated robustness in dynamic environments.

Abstract

This paper investigates payload grasping from a moving platform using a hook-equipped aerial manipulator. First, a computationally efficient trajectory optimization based on complementarity constraints is proposed to determine the optimal grasping time. To enable application in complex, dynamically changing environments, the future motion of the payload is predicted using a physics simulator-based model. The success of payload grasping under model uncertainties and external disturbances is formally verified through a robustness analysis method based on integral quadratic constraints. The proposed algorithms are evaluated in a high-fidelity physical simulator, and in real flight experiments using a custom-designed aerial manipulator platform.

Figures (8)

• Figure 1: Payload grasping from an UGV with a hook-based manipulator.
• Figure 2: Quadrotor–hook model with the three main coordinate frames.
• Figure 3: Left: 3 phases of the motion trajectory, right: grasping conditions.
• Figure 4: Left: Interconnection of nominal LTV system $G$ and uncertainty $\Delta$ with disturbance input $d$ and performance output $e$. Right: Extended LTV system with IQC filter $\Psi$. seiler_finite_2019
• Figure 5: UGV reference path (blue), and initial drone positions for robustness analysis (orange).