Scholarlink

<p>Dynamic aerial grasping remains a significant challenge due to the strong dynamic coupling between the flight platform and the manipulator, as well as the unmodeled transient shocks during contacts. Conventional methods frequently utilize explicit force sensors, which inherently involves certain additional constraints. To address these challenges, this work presents a sensorless, energy-based compliant control framework for dynamic aerial manipulation. Grounded in Hamiltonian mechanics, the proposed approach achieves robust trajectory tracking via an input-to-state stable controller. Instead of force/torque sensors, a generalized contact estimator is developed to extract external contact forces and power, effectively serving as an interaction trigger. Furthermore, a state-aware hierarchical strategy, integrating a finite state machine and a bio-inspired admittance filter, is proposed to seamlessly modulate system compliance. This allows the aerial manipulator to dynamically mitigate the impact of physical contact on the system and adaptively stiffen for stable payload transportation. The proposed framework is validated on a custom aerial manipulator. Crucially, our method leverages the contact power ramp rate as a high-fidelity signature, of which the dramatic about <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(10^2\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mn>10</mn> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation> W/s surge upon impact against a near-zero baseline effectively eradicates false-positive detections. Both simulations and real-world experiments demonstrate that the system achieves reliable pre-grasp contact detection, robust external force estimation, and smooth dynamic payload handling without the need of dedicated force/torque sensors.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Sensorless contact detection and compliant control for dynamic aerial grasping

  • 期刊论文

  • 发表日期：

    2026年7月23日

查看全文
  • Tong Li,
  • Yuanzhe Cui,
  • Lingchong Gao,
  • Wenkai Zhang,
  • Johannes Fottner,
  • Qirong Tang

摘要

Dynamic aerial grasping remains a significant challenge due to the strong dynamic coupling between the flight platform and the manipulator, as well as the unmodeled transient shocks during contacts. Conventional methods frequently utilize explicit force sensors, which inherently involves certain additional constraints. To address these challenges, this work presents a sensorless, energy-based compliant control framework for dynamic aerial manipulation. Grounded in Hamiltonian mechanics, the proposed approach achieves robust trajectory tracking via an input-to-state stable controller. Instead of force/torque sensors, a generalized contact estimator is developed to extract external contact forces and power, effectively serving as an interaction trigger. Furthermore, a state-aware hierarchical strategy, integrating a finite state machine and a bio-inspired admittance filter, is proposed to seamlessly modulate system compliance. This allows the aerial manipulator to dynamically mitigate the impact of physical contact on the system and adaptively stiffen for stable payload transportation. The proposed framework is validated on a custom aerial manipulator. Crucially, our method leverages the contact power ramp rate as a high-fidelity signature, of which the dramatic about \(10^2\) 10 2 W/s surge upon impact against a near-zero baseline effectively eradicates false-positive detections. Both simulations and real-world experiments demonstrate that the system achieves reliable pre-grasp contact detection, robust external force estimation, and smooth dynamic payload handling without the need of dedicated force/torque sensors.