Two complementary physics priors for robust in-hand rolling manipulation are introduced: a global grasp-quality prior derived from classical grasp analysis and a local contact-geometry prior based on fingertip curvature.
Abstract
In-hand manipulation without external sensing is challenging due to uncertainties from finger-object contacts and disturbances by gravity. While reinforcement learning has shown promise in learning complex finger gaiting, existing approaches do not prioritize maintaining well-conditioned grasps for sustained manipulation. We introduce two complementary physics priors for robust in-hand rolling: a global grasp-quality prior derived from classical grasp analysis and a local contact-geometry prior based on fingertip curvature. The grasp-quality prior is used as a dense reward-shaping term that encourages well-distributed contacts with improved worst-case wrench resistance. The contact-geometry prior is expressed in the fingertip geometry that mechanically shapes the contact interface toward task-aligned rolling while reducing off-axis drift. We evaluate the effect of these priors on learning in-hand rolling manipulation for a multifingered robotic hand manipulating three different objects at four palm orientations. Results show significant improvement in rotation efficiency, grasp stability, and disturbance rejection, suggesting that physics priors embedded in both learning and fingertip morphology improve task robustness and sim-to-real transfer. An overview video can be found at https://youtu.be/pdd1wHxQnJM?si=dM-U5kiiPTYsk3Pk.
In-hand manipulation allows multi-fingered dexterous hands to reconfigure grasped objects without releasing and regrasping them. This improves manipulation efficiency by reducing repeated grasp acquisition and large arm motions. However, most learning-based methods focus on reorientation, continuous rotation, or translation, whereas many tasks require joint control of object position and orientation. We formulate this capability as in-hand 6D object pose reaching: starting from an existing grasp, coordinated finger motions move the object to a palm-relative target pose. We present POISE (Palm-relative Object reaching In SE(3)), a sim-to-real reinforcement learning framework for this task. POISE combines diverse stable-grasp initialization, goal- and geometry-conditioned control, an adaptive 6D goal curriculum, and a compact reward scheme for pose reaching and grasp preservation. In simulation, diverse initialization raises held-out-grasp success from 40.1% to 51.5% and post-drop recovery from 33.8% to 72.9%; the curriculum raises full-range success from 6.2% to 59.5%. On hardware, the grasp-maintenance reward improves three-target sequence success from 20% to 80%. In real-world experiments, POISE reaches successive 6D targets without manual reset across multiple object geometries and wrist orientations, and recovers from external disturbances. To support further research in dexterous manipulation, we will release our code at https://junxiaolin.github.io/poise-website/.
Jun-Xiao Lin, Tian-Yue Wu, Jie Yin et al.· 0 citations
Quantitative experiments showed that the proposed method generally outperformed the baselines in mass and CoM variations, particularly in terms of success rate, while maintaining robust performance across the evaluated conditions.
Jinseok Kim, Iksu Choi, Hunjo Lee et al.· Intelligent Service Robotics· 0 citations
Robust three-finger grasping under physical-domain variation remains challenging because contact stability can change substantially with object mass, effective friction, and observation noise. This work develops U-GRA, a conservative offline-to-online residual adaptation framework for simulated three-finger grasping. U-GRA introduces a unified prior-preserving and critic-disagreement-regulated architecture that couples a frozen behavioral prior with a spectrally normalized and bounded residual stream, scalar Twin-Q reliability assessment, and critic-conditioned residual fusion. The framework first learns a nominal behavioral prior from successful demonstrations and then freezes it as a stable action anchor during online adaptation. Before execution, the twin critics evaluate a candidate action formed from the prior action and the bounded residual proposal, and their absolute scalar Q-value disagreement conditions a state-dependent gate that regulates residual-injection strength. Experiments are conducted in CoppeliaSim using an offline dataset of 40,000 successful demonstrations and online randomization of object mass, effective friction, and observation noise. Across three independent seeds, U-GRA achieves a mean success rate of 84.8±2.3%, a normalized return of 82.7±4.1, and a jitter value of 0.12±0.03. Relative to AWAC-Res, the strongest evaluated baseline, U-GRA improves mean success by 9.2 percentage points and reduces jitter by 57.1%. It also retains the highest mean success rate and normalized return over the unseen simulated high-mass–low-friction OOD region. These results provide simulation evidence that preserving a nominal behavioral prior while regulating bounded residual correction through critic disagreement improves three-finger grasping robustness under physical-domain variation.
Juncheng Zhu, Zhan Gao, Zhile Yang et al.· Machines· 0 citations
Although grasp prediction is well-established for rigid robotic hands, it remains particularly challenging for pneumatic soft hands due to their high compliance, nonlinear actuation, and limited sensing capabilities. To address these challenges, an integrated hardware-algorithm method is developed that leverages proprioceptive signals captured during a prelift interaction phase. A multichannel, IMU-based modular sensing system is embedded in each finger to capture joint orientation and acceleration in real time during grasp attempts. The Soft-hand Fusion Attention Network (SoftFA-Net), a lightweight multimodal model, is designed to fuse pressure, acceleration, and orientation signals, and to adaptively emphasize informative temporal segments, sensing modalities, and feature channels via a hierarchical attention mechanism. Experimental results show that the method achieves 92.9% accuracy (macro-F1 92.0%) on seen objects and yields 88.6% accuracy (macro-F1 89.0%) on unseen objects. It is further deployed in an online grasping task, enabling iterative refinement of the grasping strategy based on prediction feedback. This work provides a practical and extensible solution for pneumatic soft hand grasp prediction, laying the groundwork for future real-time multimodal sensing integration in complex manipulation tasks.
Ke-Han Ding, Rui-Chen Zhen, Ming Cheng et al.· IEEE/ASME transactions on me...· 0 citations
Dexterous manipulation with multi-fingered robot hands promises human-level dexterity, but collecting large-scale dexterous robot hand data remains difficult. Learning from human demonstrations has emerged as a scalable alternative to robot teleoperation, providing strong priors on object interaction and contact strategies. Recent sim-to-real RL methods incorporate such priors, but often (i) omit rewards that explicitly incentivize precise contact, yielding weak real-world performance, and/or (ii) generalize poorly to unseen object instances. We propose DemoMimic (Dexterous Motion Mimic), a policy that manipulates objects by focusing on their geometry local to the contact points. Its contact-centric rewards encourage precise contact and improve sim-to-real consistency, yielding a single real-world policy that transfers across objects of varying shape, scale, mass, and friction wherever local contact structure is preserved. Real-world ablations show that DemoMimic achieves 71% success across 16 objects, four tasks, and two robot-hand embodiments, with the smallest sim-to-real drop compared to baselines.
Satvik Sharma, Samrat Sahoo, Huang Huang et al.· 0 citations
In robotic manipulation studies, grasping is often treated as a binary success or failure problem, usually defined by whether the object simply stays in the hand. For forceful tool use, however, this view is insufficient because grasp compliance becomes a critical factor governing how the hand and tool behave under load. Compliance arises from coupled kinematics, grasp configuration, passive mechanics, and contact conditions, producing nonlinear behavior in which deformation and interaction forces influence each other. Understanding this relationship is essential for predictive models of how a grasped tool and a compliant hand jointly respond to external loading. In underactuated hands, these effects are amplified: such designs offer low cost and adaptive grasping, but make compliance behavior more difficult to model and predict. Our goal is therefore to develop a predictive model for grasped tool behavior during forceful interactions. To address this challenge, we introduce an analytical model informed neural network (AMINN), a hybrid predictive model that combines an analytical mechanics layer with data driven learning to estimate grasp stability and in hand tool displacement under external loading. The model is evaluated on a three finger underactuated robotic hand and shows strong predictive capability with mechanically meaningful outputs across diverse loading conditions. Compared with a black box multilayer perceptron baseline, AMINN also achieves better energy based physical consistency. Beyond prediction accuracy alone, this framework advances physically interpretable learning for robotic manipulation and supports more reliable, safer, and more trustworthy autonomous tool use in safety critical settings during forceful interaction.
Qianwen Zhao, Long Wang· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.