cs.ROSep 27, 2026

Multi-Modal Non-Prehensile Estimation of Physical Parameters via Press-and-Pull Tipping

Authors: Steven M. Hyland, Jing Xiao, Cagdas D. Onal

Organizations: Worcester Polytechnic Institute

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Recovering physical properties of unknown objects through non-prehensile interaction is challenging because no single manipulation primitive reveals all relevant parameters. Planar pushing couples mass and friction, while conventional tipping cannot recover friction and may fail entirely when low-friction or curved-base objects slide or rotate instead of tipping. We introduce a multi-modal estimation framework that combines a sliding interaction with a press-and-pull tipping primitive to recover object mass, center of mass height, and surface friction. The press-and-pull interaction increases the object-table sliding threshold and stabilizes the pivot, enabling controlled tipping without any prior geometric object model. Wrist force/torque sensing, RGB-D perception, and robot proprioception are fused to estimate the physical parameters from the two complementary interaction modes. Experiments on an ABB IRB120 across four objects with varied geometry, mass, center of mass, and friction achieve low relative error, while successfully operating on curved-base objects that fail under conventional forward tipping. The results demonstrate that complementary non-prehensile interactions can recover a compact set of physical parameters without grasping, a prior object model, or learned interaction dynamics.

Figures & tables

Figure 1: The press-and-pull primitive. A downward press raises the sliding threshold at the object–table interface, while a horizontal pull induces tipping about the near-robot pivot edge.
Figure 2: Adaptive press procedure on the monitor object. (right) Slip is detected by deviation from the arc motion and observed vs. expected object pose. Press force that is ultimately found remains constant through the arc for all object angles θ\theta (from proprioception).
Figure 3: Multi-modal pipeline. Pose estimation is used to detect finger slip and assist with contact-point selection. Friction estimation relies on output from NLS (non-linear least squares) estimation.
Figure 4: Heart object during the press-and-pull mode. The finger first applies a downward press force. Upon reaching the press force, the finger retracts and applies a pull force, which continues while the finger maintains the press force throughout tipping. This continues until the measured tangential force falls to the defined safety margin ηsafety\eta_{safety} .
Figure 5: (left) Robot finger used for both interaction modes. High-friction spherical tip provides better traction during press-and-pull and reduces required press force magnitude. (right) four objects tested, varying in size, shape, weight, friction, and CoM.
Figure 6: Pointcloud and contact point selection. Candidate points taken from the pointcloud near the extremes of object height, and selected according to feasibility and the heuristics defined in Sec. 3.2 .
ConditionRequirementServes
VisibilityTop face and pivot edge each seen by a cameraContact selection
ReachabilityTop contact and pull arc within the workspacePress-and-pull
Top frictionHolds finger no-slip under Fpress\mathbf{F}_{press}Adaptive press
RigidityPivot-to-CoM moment arm fixed in {O}\{O\}Torque balance
SlideabilitySlides before tipping under a low pushMode 1 ( μt\mu_{t} )
Table 1: Admissible object class: conditions an object must satisfy, and the component of the framework each one serves.
 mm (kg)zcz_{c} (cm)μt\mu_{t}
ObjectGTEstimateGTEstimateGTEstimate
Box0.6760.707±0.0120.707\pm 0.01215.0015.10±0.2815.10\pm 0.280.1560.149±0.0030.149\pm 0.003
Heart0.2390.242±0.0060.242\pm 0.00610.0011.51±0.2711.51\pm 0.270.2560.253±0.0060.253\pm 0.006
Flashlight0.3870.396±0.0040.396\pm 0.0049.389.38±0.129.38\pm 0.120.2400.234±0.0020.234\pm 0.002
Monitor5.0405.274±0.0085.274\pm 0.00823.2024.42±0.0524.42\pm 0.050.6430.614±0.0010.614\pm 0.001
Table 2: Estimation results on physical objects. m^\hat{m} and z^c\hat{z}_{c} are the mean ±\pm one SD over N=10N\text{=}10 press-and-pull trials; μ^t\hat{\mu}_{t} follows from a single Mode 1 push and the recovered mass.
Figure 7: Ground-truth and estimated parameter values across the four objects; error bars show ±1\pm 1 SD.
Figure 8: Relative error by object and parameter; error bars show ±1\pm 1 SD.
More in this topic

Non-Prehensile Robotic Manipulation

See all 16 papers

Explore similar work

CardsList
Sep 11, 2026

Before the Tipping Point: Force-Guided Active Perception for Shape-Agnostic Estimation of 3D Centers of Mass

Estimating the 3D center of mass of unknown objects is challenging when grasping is infeasible, geometry is irregular, or mass distribution is uneven. We present a force-based method that estimates CoM height and mass from a single sub-critical tipping experiment by a robot manipulator. The robot applies a quasistatic elevated push and retract motion, using force-angle measurements recorded during tipping to identify parameters from the object trajectory. Our proposed push-retract cycle mitigates frictional bias, enabling generalized fitting. We experimentally validate our method using a robot manipulator with a six-axis force torque sensor on varying types of objects without prior shape information and without specific models. We also propose a method to prevent toppling, keeping the object in a sub-critical tipping regime by leveraging a safety margin. In experimental studies, our method recovers mass, CoM height, and toppling angle with relative errors below 5.0 percent across all unknown objects. This work demonstrates reliable 3D inertial parameter estimation under proper safety thresholds in tipping. Our proposed method informs and enables reliable non-prehensile manipulation and robotic grasping of challenging objects that were previously infeasible.

Steven M. Hyland, Jing Xiao, Cagdas D. Onal

Robotics Engineering Department, Worcester Polytechnic Institute, MA 01609, USA

Active PerceptionForce-Controlled Robotic Manipulation
AskPDF
May 25, 2026

PhyPush: One Push is All You Need for Sensorless Physical Property Estimation with Physics-Guided Transformers

Accurately estimating object mass and friction is fundamental to reliable robotic manipulation. While interactive perception is powerful, most approaches rely on specialized hardware like force/torque sensors, limiting scalability. This paper introduces PhyPush, a physics-guided Transformer that estimates an object's mass and friction coefficient using only end-effector velocity from a single push, data readily available on standard robotic arms. By incorporating Newton's second law and the Coulomb friction model through a physics-guided loss, the model improves physical consistency and generalizes to unseen objects and surfaces. Across diverse setups, PhyPush consistently achieves highly accurate estimations in challenging out-of-domain conditions. In simulation, it reduces error by over 10% compared to a baseline with privileged force data, while in real-world experiments, it successfully zero-shot transfers from simulation to outperform a purely data-driven baseline.

Koyo Fujii, Luis Figueredo, Praminda Caleb-Solly +4

School of Computer Science, University of Nottingham, Nottingham, United Kingdom · Dept. of Industrial Engineering, University of Bologna, Bologna, Italy

Physical ReasoningRobotic ManipulationSystem IdentificationPhysics-Informed ML
AskPDF
Mar 12, 2026

Push, Press, Slide: Mode-Aware Planar Contact Manipulation via Reduced-Order Models

Non-prehensile planar manipulation, including pushing and press-and-slide, is critical for diverse robotic tasks, but notoriously challenging due to hybrid contact mechanics, under-actuation, and asymmetric friction limits that traditionally necessitate computationally expensive iterative control. In this paper, we propose a mode-aware framework for planar manipulation with one or two robotic arms based on contact topology selection and reduced-order kinematic modeling. Our core insight is that complex wrench-twist limit surface mechanics can be abstracted into a discrete library of physically intuitive models. We systematically map various single-arm and bimanual contact topologies to simple non-holonomic formulations, e.g. unicycle for simplified press-and-slide motion. By anchoring trajectory generation to these reduced-order models, our framework computes the required object wrench and distributes feasible, friction-bounded contact forces via a direct algebraic allocator. We incorporate manipulator kinematics to ensure long-horizon feasibility and demonstrate our fast, optimization-free approach in simulation across diverse single-arm and bimanual manipulation tasks. Supplementary videos and additional information are available at: https://sites.google.com/view/pushpressslide

Melih Özcan, Umut Orguner, Ozgur S. Oguz

Dept. of Electrical and Electronics Engineering, Middle East Technical University, Ankara, Turkiye. · Dept. of Computer Engineering, Bilkent University, Ankara, Turkiye.

Non-Prehensile Robotic ManipulationRobotic ManipulationContact-Rich Robotic ManipulationMotion PlanningBimanual Robotic ManipulationReduced-Order Modeling
AskPDF