arXiv subjects

Yuki Noguchi

Publications and source records attributed to Yuki Noguchi.

17 recordsLinked to original sources

arXiv2026

YUBI: Yielding Universal Bidigital Interface for Bimanual Dexterous Manipulation at Scale

We introduce Yielding Universal Bidigital Interface (YUBI), a finger-aligned gripper designed to enable intuitive, ergonomic, and scalable data collection for bimanual dexterous manipulation. While handheld data collection systems such as Universal Manipulation Interface (UMI) enable affordable data collection, their bulky pistol-grip designs can pose ergonomic and usability challenges for fine-grained, dexterous manipulation tasks. To address this, YUBI presents a distinct design principle: yielding, finger-driven actuation that directly maps human finger movements to gripper jaw motion. Using the YUBI devices, we set up a data collection system with integrated VR-based 6 DoF tracking of the gripper, ensuring high-fidelity trajectory data acquisition. We curate a UMI-based dataset of unprecedented scale: 8,434 hours across 1.20M episodes and 119 tasks. Experiments show that YUBI offers advantages over the UMI gripper in versatility for complex bimanual tasks, dexterity, and operational efficiency. A single policy trained on the YUBI dataset transfers across multiple bimanual robots (UR, Franka, and ELEY) simply by mounting the gripper on each platform, confirming that the collected data are directly executable as policy supervision. We release the gripper hardware, data-collection software, and dataset as one integrated stack, offering the open community a reproducible path to large-scale data acquisition for advancing robotic foundation models.

cs.RO↗

arXiv2026

A characterization of terminal planar networks by forbidden structures

The class of terminal planar networks was recently introduced from a biological perspective in relation to the visualization of phylogenetic networks, and its connection to upward planar networks has been established. We provide a Kuratowski-type theorem that characterizes terminal planar networks by a finite set of forbidden structures, defined via six families of 0/1-labeled graphs. Another characterization based on planarity of supergraphs yields linear-time algorithms for testing terminal planarity and for computing such planar drawings. We describe an application that is potentially relevant in broader, non-phylogenetic settings. We also discuss a connection of our main result to an open problem on the forbidden structures of single-source upward planar networks.

math.CO↗

arXiv2025

AIRoA MoMa Dataset: A Large-Scale Hierarchical Dataset for Mobile Manipulation

As robots transition from controlled settings to unstructured human environments, building generalist agents that can reliably follow natural language instructions remains a central challenge. Progress in robust mobile manipulation requires large-scale multimodal datasets that capture contact-rich and long-horizon tasks, yet existing resources lack synchronized force-torque sensing, hierarchical annotations, and explicit failure cases. We address this gap with the AIRoA MoMa Dataset, a large-scale real-world multimodal dataset for mobile manipulation. It includes synchronized RGB images, joint states, six-axis wrist force-torque signals, and internal robot states, together with a novel two-layer annotation schema of sub-goals and primitive actions for hierarchical learning and error analysis. The initial dataset comprises 25,469 episodes (approx. 94 hours) collected with the Human Support Robot (HSR) and is fully standardized in the LeRobot v2.1 format. By uniquely integrating mobile manipulation, contact-rich interaction, and long-horizon structure, AIRoA MoMa provides a critical benchmark for advancing the next generation of Vision-Language-Action models. The first version of our dataset is now available at https://huggingface.co/datasets/airoa-org/airoa-moma .

cs.RO↗

arXiv2025

A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation

Robot manipulation has seen tremendous progress in recent years, with imitation learning policies enabling successful performance of dexterous and hard-to-model tasks. Concurrently, scaling data and model size has led to the development of capable language and vision foundation models, motivating large-scale efforts to create general-purpose robot foundation models. While these models have garnered significant enthusiasm and investment, meaningful evaluation of real-world performance remains a challenge, limiting both the pace of development and inhibiting a nuanced understanding of current capabilities. In this paper, we rigorously evaluate multitask robot manipulation policies, referred to as Large Behavior Models (LBMs), by extending the Diffusion Policy paradigm across a corpus of simulated and real-world robot data. We propose and validate an evaluation pipeline to rigorously analyze the capabilities of these models with statistical confidence. We compare against single-task baselines through blind, randomized trials in a controlled setting, using both simulation and real-world experiments. We find that multi-task pretraining makes the policies more successful and robust, and enables teaching complex new tasks more quickly, using a fraction of the data when compared to single-task baselines. Moreover, performance predictably increases as pretraining scale and diversity grows. Project page: https://toyotaresearchinstitute.github.io/lbm1/

cs.RO↗

arXiv2024

Robust Imitation Learning for Mobile Manipulator Focusing on Task-Related Viewpoints and Regions

We study how to generalize the visuomotor policy of a mobile manipulator from the perspective of visual observations. The mobile manipulator is prone to occlusion owing to its own body when only a single viewpoint is employed and a significant domain shift when deployed in diverse situations. However, to the best of the authors' knowledge, no study has been able to solve occlusion and domain shift simultaneously and propose a robust policy. In this paper, we propose a robust imitation learning method for mobile manipulators that focuses on task-related viewpoints and their spatial regions when observing multiple viewpoints. The multiple viewpoint policy includes attention mechanism, which is learned with an augmented dataset, and brings optimal viewpoints and robust visual embedding against occlusion and domain shift. Comparison of our results for different tasks and environments with those of previous studies revealed that our proposed method improves the success rate by up to 29.3 points. We also conduct ablation studies using our proposed method. Learning task-related viewpoints from the multiple viewpoints dataset increases robustness to occlusion than using a uniquely defined viewpoint. Focusing on task-related regions contributes to up to a 33.3-point improvement in the success rate against domain shift.

cs.RO↗

arXiv2023

Level set-based shape optimization of deformable structures for manipulating sound propagation

In this paper, we propose a level set-based shape optimization method for acoustic wave propagation problems with a deformable structure. First, we propose a mathematical model for acoustic wave propagation with a deformed structure based on coordinate transformation and the Eulerian approach. Next, we formulate the shape optimization problem and perform sensitivity analysis based on the shape derivative concept. We then construct an optimization algorithm based on the framework of a level set-based shape and topology optimization method. Finally, we provide two-dimensional optimization examples that demonstrate the effectiveness of our proposed method in providing optimized designs with desired functionality, while considering the structural deformation.

cs.CE↗

arXiv2023

Two-Phase Topology Optimization for Metamaterials with Negative Poisson's Ratio

Although recent developments in 3D printing technology have made it possible to fabricate metamaterials with characteristic mechanical properties, it is not easy to fabricate complex shapes containing cavities. In this study, a composite structure comprising two types of materials without a cavity was optimized. Moreover, a mechanical metamaterial with a negative Poisson's ratio that can be fabricated using an additive manufacturing method was developed. First, a homogenization method that characterizes the properties of composite structures was briefly described. Then, an optimization problem to realize a negative Poisson's ratio was formulated, and a level set-based topology optimization method was proposed to solve the abovementioned problem. Next, three-dimensional numerical examples are presented to confirm the effectiveness of the proposed method, and the deformation behaviors of the optimized designs are numerically examined. Furthermore, a sample containing the optimized design with negative Poisson's ratio for the tensile test was fabricated using a 3D printer. We conducted some experiments to evaluate its mechanical performance.

physics.app-ph↗

arXiv2023

Level set-based multiscale topology optimization for a thermal cloak design problem using the homogenization method

Artificially designed composite materials consist of microstructures, that exhibit various thermal properties depending on their shapes, such as anisotropic thermal conductivity. One of the representative applications of such composite materials for thermal control is the thermal cloak. This study proposed a topology optimization method based on a level set method for a heat conduction problem to optimally design composite materials that achieve a thermal cloak. The homogenization method was introduced to evaluate its effective thermal conductivity coefficient. Then, we formulated a multiscale topology optimization method for the composite materials in the framework of the homogenization method, where the microstructures were optimized to minimize objective functions defined using the macroscopic temperature field. We presented examples of optimal structures in a two-dimensional problem and discussed the validity of the obtained structures.

cs.CE↗

arXiv2023

Exceptional points in cylindrical elastic media with radiation loss

Exceptional points (EPs) are singular points on a parameter space at which some eigenvalues (scattering poles) and their corresponding eigenmodes coalesce. This study shows the existence of second- and third-order EPs in cylindrical elastic systems with radiation loss. We consider multilayered cylindrical solids under the plane-strain condition placed in a background elastic or acoustic medium. Elastic and acoustic waves propagating in the background media are subject to the radiation loss. We optimize the radii and the material constants of the multilayered solids, such that some scattering poles coalesce on the complex frequency plane. Some numerical experiments are performed to confirm that the coalescence originates from EPs. We expect that this study provides a new approach for enhancing mechanical sensors.

physics.class-ph↗

arXiv2023

Optimal design of compliant displacement magnification mechanisms using stress-constrained topology optimization based on effective energy

In this paper, stress-constrained topology optimization is applied to the design of compliant displacement magnification mechanisms. By formulating the objective function based on the concept of effective energy, it is not necessary to place artificial spring components at the boundaries of the output and input ports as in previous methods. This makes it possible to design mechanisms that do not receive a reaction force at the output port, such as sensors. Furthermore, by imposing a constraint on the maximum stress evaluated in terms of the p-norm of the von Mises equivalent stress, problems such as stress concentration can be avoided. Several numerical examples of displacement magnification mechanisms are provided to demonstrate the effectiveness of the proposed method.

cs.CE↗

arXiv2022

World Robot Challenge 2020 -- Partner Robot: A Data-Driven Approach for Room Tidying with Mobile Manipulator

Tidying up a household environment using a mobile manipulator poses various challenges in robotics, such as adaptation to large real-world environmental variations, and safe and robust deployment in the presence of humans.The Partner Robot Challenge in World Robot Challenge (WRC) 2020, a global competition held in September 2021, benchmarked tidying tasks in the real home environments, and importantly, tested for full system performances.For this challenge, we developed an entire household service robot system, which leverages a data-driven approach to adapt to numerous edge cases that occur during the execution, instead of classical manual pre-programmed solutions. In this paper, we describe the core ingredients of the proposed robot system, including visual recognition, object manipulation, and motion planning. Our robot system won the second prize, verifying the effectiveness and potential of data-driven robot systems for mobile manipulation in home environments.

cs.RO↗

arXiv2022

Topology optimization for acoustic structures considering viscous and thermal boundary layers using a sequential linearized Navier-Stokes model

This study proposes a level set-based topology optimization method for designing acoustic structures with viscous and thermal boundary layers in perspective. Acoustic waves propagating in a narrow channel are damped by viscous and thermal boundary layers. To estimate these viscothermal effects, we first introduce a sequential linearized Navier-Stokes model based on three weakly coupled Helmholtz equations for viscous, thermal, and acoustic pressure fields. Then, the optimization problem is formulated, where a sound-absorbing structure comprising air and an isothermal rigid medium is targeted, and its sound absorption coefficient is set as an objective function. The adjoint variable method and the concept of the topological derivative are used to approximately obtain design sensitivity. A level set-based topology optimization method is used to solve the optimization problem. Two-dimensional numerical examples are provided to support the validity of the proposed method. Moreover, the mechanisms that lead to the high absorption coefficient of the optimized design are discussed.

cs.CE↗

arXiv2022

Topology optimization with a closed cavity exclusion constraint for additive manufacturing based on the fictitious physical model approach

This paper proposes a topology optimization method that considers the geometric constraint of no closed cavities to improve the effectiveness of additive manufacturing based on the fictitious physical model approach. First, the basic topology optimization concept and level set-based method are introduced. Next, the fictitious physical model for a geometric constraint in the topology optimization framework is discussed. Then, a model for the geometric constraint of no closed cavities for additive manufacturing is proposed. Numerical examples are provided to validate the proposed model. In addition, topology optimization considering the geometric constraint is formulated, and topology optimization algorithms are constructed using the finite element method. Finally, optimization examples are provided to validate the proposed topology optimization method.

cs.SE↗

arXiv2022

A coupled-mode theory for two-dimensional exterior Helmholtz problems based on the Neumann and Dirichlet normal mode expansion

This study proposes a novel coupled-mode theory for two-dimensional exterior Helmholtz problems. The proposed approach is based on the separation of the entire space R2 into a fictitious disk and its exterior. The disk is allocated in such a way that it comprises all the inhomogeneity; therefore, the exterior supports cylindrical waves with a continuous spectrum. For the interior, we expand an unknown wave field using normal modes that satisfy some auxiliary boundary conditions on the surface of the disk. For the interior expansion, we propose combining the Neumann and Dirichlet normal modes. We show that the proposed expansion sacrifices L2 orthogonality but significantly improve the convergence. Finally, we present some numerical verifications of the proposed coupled-mode theory.

math.NA↗

arXiv2021

Tool as Embodiment for Recursive Manipulation

Humans and many animals exhibit a robust capability to manipulate diverse objects, often directly with their bodies and sometimes indirectly with tools. Such flexibility is likely enabled by the fundamental consistency in underlying physics of object manipulation such as contacts and force closures. Inspired by viewing tools as extensions of our bodies, we present Tool-As-Embodiment (TAE), a parameterization for tool-based manipulation policies that treat hand-object and tool-object interactions in the same representation space. The result is a single policy that can be applied recursively on robots to use end effectors to manipulate objects, and use objects as tools, i.e. new end-effectors, to manipulate other objects. By sharing experiences across different embodiments for grasping or pushing, our policy exhibits higher performance than if separate policies were trained. Our framework could utilize all experiences from different resolutions of tool-enabled embodiments to a single generic policy for each manipulation skill. Videos at https://sites.google.com/view/recursivemanipulation

cs.RO↗

arXiv2021

Extended level set method: a multiphase representation with perfect symmetric property, and its application to multi material topology optimization

This paper provides an extended level set (X-LS) based topology optimiza- tion method for multi material design. In the proposed method, each zero level set of a level set function ϕij represents the boundary between materials i and j. Each increase or decrease of ϕij corresponds to a material change between the two materials. This approach reduces the dependence of the initial configuration in the optimization calculation and simplifies the sensitivity analysis. First, the topology optimization problem is formulated in the X-LS representation. Next, the reaction-diffusion equation that updates the level set function is introduced, and an optimization algorithm that solves the equilibrium equations and the reaction-diffusion equation using the fi- nite element method is constructed. Finally, the validity and utility of the proposed topology optimization method are confirmed using two- and three- dimensional numerical examples.

cs.CE↗

arXiv2021

Topology optimization of acoustic metasurfaces by using a two-scale homogenization method

In this paper, we propose a level set-based topology optimization method for the unit-cell design of acoustic metasurfaces by using a two-scale homogenization method. Based on previous works, we first propose a homogenization method for acoustic metasurfaces that can be combined with topology optimization. In this method, a nonlocal transmission condition depending on the unit cell of the metasurface appears in a macroscale problem. Next, we formulate an optimization problem within the framework of a level set-based topology optimization method, wherein an objective functional is expressed as the macroscopic responses obtained through the homogenization, and material distributions in the unit cell are set as design variables. A sensitivity analysis is conducted based on the concept of the topological derivative. To confirm the validity of the proposed method, two-dimensional numerical examples are provided. First, we provide a numerical example that supports the validity of the homogenization method, and we then perform optimization calculations based on the waveguide settings of the acoustic metasurfaces. In addition, we discuss the mechanism of the obtained optimized structures.

cs.CE↗