Selected work

Representative Papers

Estimating Force Interactions of Deformable Linear Objects from their Shapes

Feb 01, 2026

This work addresses the challenge of accurately estimating external forces applied at non-end-effector locations on deformable linear objects—such as wires—during robotic manipulation, a key limitation for safe physical interaction. The authors propose a novel method that relies solely on depth-camera–derived shape observations, eliminating the need for additional force sensors or assumptions about end-effector contact. Under static equilibrium conditions, the approach formulates a linear system based on force–moment consistency to analytically estimate both the location and magnitude of external forces. To the best of the authors’ knowledge, this is the first technique capable of directly inferring arbitrary external forces from purely visual shape data, making it suitable for scenarios involving indirect manipulation or passive obstacles. Extensive simulations and real-world experiments demonstrate the method’s high accuracy and robustness in localizing and quantifying external forces.

0 citationsRead paper

Robotic Manipulation of a Rotating Chain with Bottom End Fixed

Jun 23, 2025

This work addresses the problem of robotic manipulator control for stabilizing and shaping a uniform flexible chain—such as a drill string or filament—rigidly fixed at one end and rotating uniformly about its base. To overcome the challenge of reliable transition from static to rotational configurations, we first establish that the chain’s configuration space is homeomorphic to a three-dimensional cube; leveraging this topological insight, we devise a mode-switching strategy that jointly ensures stability and kinematic feasibility. Integrating differential-geometric modeling with motion planning, we analytically characterize the chain’s rotational dynamics and design a closed-loop controller operating directly in configuration space. Experimentally, we achieve, for the first time, repeatable and stable transitions from the static state to both the first and second rotational modes. These results validate the proposed framework’s effectiveness and practicality in ensuring operational safety and efficiency.

0 citationsRead paper

Recent publications

Latest Papers

Estimating Force Interactions of Deformable Linear Objects from their Shapes

Feb 01, 2026

This work addresses the challenge of accurately estimating external forces applied at non-end-effector locations on deformable linear objects—such as wires—during robotic manipulation, a key limitation for safe physical interaction. The authors propose a novel method that relies solely on depth-camera–derived shape observations, eliminating the need for additional force sensors or assumptions about end-effector contact. Under static equilibrium conditions, the approach formulates a linear system based on force–moment consistency to analytically estimate both the location and magnitude of external forces. To the best of the authors’ knowledge, this is the first technique capable of directly inferring arbitrary external forces from purely visual shape data, making it suitable for scenarios involving indirect manipulation or passive obstacles. Extensive simulations and real-world experiments demonstrate the method’s high accuracy and robustness in localizing and quantifying external forces.

0 citationsRead paper

Robotic Manipulation of a Rotating Chain with Bottom End Fixed

Jun 23, 2025

This work addresses the problem of robotic manipulator control for stabilizing and shaping a uniform flexible chain—such as a drill string or filament—rigidly fixed at one end and rotating uniformly about its base. To overcome the challenge of reliable transition from static to rotational configurations, we first establish that the chain’s configuration space is homeomorphic to a three-dimensional cube; leveraging this topological insight, we devise a mode-switching strategy that jointly ensures stability and kinematic feasibility. Integrating differential-geometric modeling with motion planning, we analytically characterize the chain’s rotational dynamics and design a closed-loop controller operating directly in configuration space. Experimentally, we achieve, for the first time, repeatable and stable transitions from the static state to both the first and second rotational modes. These results validate the proposed framework’s effectiveness and practicality in ensuring operational safety and efficiency.

0 citationsRead paper