arXivSub Start free trial

IROS 2024 Papers — Page 16

IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers

TRX-Hand5: An Anthropomorphic Hand with Integrated Tactile Feedback for Grasping and Manipulation in Human Environments

Sicheng Yang, Yu Zheng

Robotic Intelligence

🎯 What it does: Designed and implemented the TRX-Hand5 humanoid hand with 13 degrees of freedom, over 1000 tactile sensing elements, and integrated position encoders and cable tension sensors to achieve tactile and proprioceptive sensing.

Tube-GAN: A Novel Virtual Tube Generation Method for Unmanned Aerial Swarms Based on Generative Adversarial Network

Shixun Zhai, Qianyi Fu

GenerationAutonomous DrivingOptimizationComputational EfficiencyGenerative Adversarial NetworkImage

🎯 What it does: Propose the Tube-GAN model, which uses GAN to generate virtual pipelines for UAV formations in dense obstacle environments, converting the optimization problem into an image generation task.

Tunable Stiffness Glove for Tremor Suppression Based on 3D Printed Structured Fabrics

Yu Chen, Yifan Wang

Time SeriesBiomedical Data

🎯 What it does: A tunable stiffness glove was developed to suppress wrist tremors in patients with Parkinson's disease and essential tremor, and its effectiveness was evaluated through three-point bending tests and human trials.

TURTLMap: Real-time Localization and Dense Mapping of Low-texture Underwater Environments with a Low-cost Unmanned Underwater Vehicle

Jingyu Song, Katherine A. Skinner

Robotic IntelligenceSimultaneous Localization and MappingPoint Cloud

🎯 What it does: Propose a low-cost, real-time localization and dense mapping method (TURTLMap) suitable for low-texture underwater environments.

Two Teachers Are Better Than One: Leveraging Depth In Training Only For Unsupervised Obstacle Segmentation

Sungmin Eum, Andre Harrison

SegmentationDepth EstimationAutonomous DrivingKnowledge DistillationImage

🎯 What it does: Proposes an unsupervised obstacle segmentation architecture based on Relation Distillation, combining a depth information teacher to enhance segmentation performance.

Two-stage pose optimization algorithm using color information for underwater SLAM with light-sectioning-based 3D scanning method

Takaki Ikeda, Hiroshi Kawasaki

Pose EstimationOptimizationSimultaneous Localization and MappingImagePoint Cloud

🎯 What it does: Proposed a two-stage pose optimization algorithm for underwater SLAM based on light stripe scanning, utilizing color information to improve localization accuracy.

U-BEV: Height-aware Bird’s-Eye-View Segmentation and Neural Map-based Relocalization

Andrea Boscolo Camiletto, Abel Gawel

SegmentationAutonomous DrivingConvolutional Neural NetworkSimultaneous Localization and MappingImageMultimodalityPoint Cloud

🎯 What it does: Proposed the U-BEV structure, achieving multi-height layer bird's eye view (BEV) semantic segmentation and integrating neural maps for localization.

UMAD: University of Macau Anomaly Detection Benchmark Dataset

Dong Li (University of Macau), Hui Kong (University of Macau)

Anomaly DetectionImageBenchmark

🎯 What it does: Proposed a reference-based anomaly detection benchmark dataset named UMAD specifically for robot patrolling scenarios, and evaluated the baseline ADr model on this dataset.

Uncertainty-aware Deep Imitation Learning and Deployment for Autonomous Navigation through Crowded Intersections

Zeyu Zhu, Huijing Zhao

Autonomous DrivingSafty and Privacy

🎯 What it does: Proposes a heterogeneous uncertainty quantification method based on imitation learning, and designs a strategy deployment scheme based on estimated uncertainty, bridging the data-driven layer with the rule-based fallback layer to achieve safe navigation at crowded intersections.

Uncertainty-Aware Deployment of Pre-trained Language-Conditioned Imitation Learning Policies

Bo Wu, Nikolai Matni

TransformerReinforcement LearningText

🎯 What it does: Proposes an uncertainty-based deployment method for pre-trained language-conditioned imitation learning agents, calibrating the model using temperature scaling and achieving uncertainty-aware decision-making through the aggregation of local information from candidate actions.

Uncertainty-Aware Semi-Supervised Semantic Key Point Detection via Bundle Adjustment

Kai Li, Shiyu Zhao

Pose Estimation

🎯 What it does: Proposed a method for jointly estimating a semantic keypoint detection model and 6DoF camera pose.

Under-actuated Robotic Gripper with Multiple Grasping Modes Inspired by Human Finger

Jihao Li, Huixu Dong

Robotic Intelligence

🎯 What it does: Designed and implemented a three-fingered, link-based variable multimodal gripper that achieves retractable and reconfigurable fingers through a single motor for multiple grasping modes;

Understanding How a 3-dimensional ZMP Exactly Decouples the Horizontal and Vertical Dynamics of the CoM-ZMP Model

Yuki Onishi, S. Kajita

Robotic IntelligencePhysics Related

🎯 What it does: Investigated how 3D ZMP precisely decouples the horizontal and vertical dynamics of the CoM-ZMP model

Understanding Robot Minds: Leveraging Machine Teaching for Transparent Human-Robot Collaboration Across Diverse Groups

Suresh Kumaar Jayaraman, H. Admoni

Explainability and InterpretabilityRobotic IntelligenceContrastive Learning

🎯 What it does: Developed and evaluated a machine teaching algorithm for teams with diverse learning capabilities, utilizing team belief representations to achieve transparency and efficiency in human-machine collaboration.

Underwater Hyperspectral Imaging for Measuring Seafloor Reflectance

Hongjie Zhang, Stefan Williams

RestorationImagePhysics Related

🎯 What it does: Introduces a method that combines underwater hyperspectral imaging, illumination formation models, and Structure from Motion (SfM) to estimate intrinsic optical properties of underwater environments and correct seabed reflectance based on radiometric measurements.

Unified Control Framework for Real-Time Interception and Obstacle Avoidance of Fast-Moving Objects with Diffusion Variational Autoencoder

Apan Dastider, Mingjie Lin

Robotic IntelligenceDiffusion modelAuto Encoder

🎯 What it does: Proposed and verified a unified real-time interception and obstacle avoidance framework, utilizing a diffusion variational autoencoder for motion planning and EKF for real-time tracking, enabling a 7-DoF robotic arm to rapidly intercept targets and avoid obstacles in dynamic environments.

UNO Push: Unified Nonprehensile Object Pushing via Non-Parametric Estimation and Model Predictive Control

Gaotian Wang, Kaiyu Hang

OptimizationRobotic Intelligence

🎯 What it does: Propose a unified framework that integrates system modeling, action generation, and control to achieve non-grasping pushing operations

Unsupervised 3D Part Decomposition via Leveraged Gaussian Splatting

Jaegoo Choy, Songhwai Oh

SegmentationGaussian SplattingVideo

🎯 What it does: Proposes an unsupervised motion-driven 3D part decomposition method based on monocular video.

Unsupervised Multiple Proactive Behavior Learning of Mobile Robots for Smooth and Safe Navigation

Arthicha Srisuchinnawong, P. Manoonpong

Robotic Intelligence

🎯 What it does: Propose a model-free neural control architecture combined with an online multi-ple active behavior learning (MPL) module, enabling mobile robots to rapidly learn and balance active behaviors such as smooth motion and collision avoidance under unsupervised, few-shot conditions.

User-customizable Shared Control for Robot Teleoperation via Virtual Reality

Rui Luo, T. Padır

Robotic Intelligence

🎯 What it does: Propose a system that allows operators to customize shared control arbitration parameters through a virtual reality (VR) interface, validated in a robot arm's buzz wire game.

Using Augmented Reality in Human-Robot Assembly: A Comparative Study of Eye-Gaze and Hand-Ray Pointing Methods

S. Tadeja, F. Forni

Robotic Intelligence

🎯 What it does: A comparative study on the usability of eye gaze and hand ray pointing interaction methods in AR interfaces for collaborative assembly tasks

Using Graphs of Convex Sets to Guide Nonconvex Trajectory Optimization

D. V. Wrangel, Russ Tedrake

OptimizationGraph

🎯 What it does: Study how to combine Graph Convex Sets (GCS) with non-convex trajectory optimization to address the non-convexity in collision avoidance trajectory planning.

Using Hip Assisted Running Exoskeleton with Impact Isolation Mechanism to Improve Energy Efficiency

Ziqi Wang, Yanhe Zhu

Anomaly DetectionRobotic Intelligence

🎯 What it does: Designed and tested a hip-assisted running exoskeleton (HARE), employing an active-passive combination constant force suspension system (CFS) and flexible structure design to reduce impact during running, further equipped with joint torque generation and abnormal gait recognition safety control strategies.

UW-SDF: Exploiting Hybrid Geometric Priors for Neural SDF Reconstruction from Underwater Multi-view Monocular Images

Zeyu Chen, Xiu Li

SegmentationNeural Radiance FieldImage

🎯 What it does: Based on the neural SDF framework, reconstructs target objects from multi-view monocular underwater images using hybrid geometric priors, and proposes a few-shot multi-view target segmentation strategy.

UWB-Based Localization System Considering Antenna Anisotropy and NLOS/Multipath Conditions

Taekyun Kim, Dongjun Lee

Simultaneous Localization and MappingPhysics Related

🎯 What it does: Researched and implemented a localization system based on UWB ranging error models, solving problems of antenna anisotropy, NLOS, and multipath interference, and including anchor self-calibration and state estimation using iterative Kalman filtering;

V-PRISM: Probabilistic Mapping of Unknown Tabletop Scenes

Herbert Wright, Tucker Hermans

SegmentationSimultaneous Localization and MappingImagePoint Cloud

🎯 What it does: Utilizing segmented RGB-D images, construct probabilistic 3D segmentation maps of desktop scenes, including occupancy rates, segmentation information, and uncertainty assessment.

V2I-Calib: A Novel Calibration Approach for Collaborative Vehicle and Infrastructure LiDAR Systems

Qianxin Qu, Shichun Guo

Autonomous DrivingPoint Cloud

🎯 What it does: Proposed a real-time robust calibration method for vehicle-infrastructure collaborative LiDAR systems based on perceived target spatial association information

V3D-SLAM: Robust RGB-D SLAM in Dynamic Environments with 3D Semantic Geometry Voting

T. Dang, M. Huber

Simultaneous Localization and MappingPoint CloudBenchmark

🎯 What it does: Proposes a robust RGB-D SLAM method called V3D-SLAM, which removes moving objects in dynamic environments through a two-stage lightweight re-evaluation process.

Valuing Attrition in a Fleet of Robots Used as Path-Based Sensors for Gathering Information in a Communications Restricted Environment

Loy McGuire, Donald Sofge

OptimizationRobotic Intelligence

🎯 What it does: Propose a new algorithm for information gathering by a limited number of robot swarms in hazardous, communication-restricted environments. The algorithm uses robot survival status along search paths as sensor events, updates target and hazard belief maps via Bayesian inference, constructs paths maximizing expected information gain, while accounting for the anticipated decline in future information collection after robot losses.

VANP: Learning Where to See for Navigation with Self-Supervised Vision-Action Pre-Training

Mohammad Nazeri, Xuesu Xiao

Autonomous DrivingTransformerVision-Language-Action ModelImage

🎯 What it does: Proposed a self-supervised visual-action model (VANP) for visual navigation pre-training, which learns to focus only on visual regions relevant to navigation tasks.

Vehicle Trajectory Prediction with Soft Behavior Constraints

Ke Ye, Nanning Zheng

Autonomous DrivingPoint Cloud

🎯 What it does: Propose a trajectory prediction framework with soft vehicle behavior constraints, integrating it as a lightweight plugin into five mainstream trajectory prediction models.

Versatile Locomotion Skills for Hexapod Robots

Tomson Qu, Tingnan Zhang

Knowledge DistillationRobotic IntelligenceReinforcement LearningSimultaneous Localization and MappingImage

🎯 What it does: Trained and deployed a six-legged robot equipped with a depth camera and visual inertial odometry (VIO), which learned three tasks in a simulated environment: climbing stairs, obstacle avoidance, and navigating under obstacles such as tables. Subsequently, the model achieved sim-to-real transfer, demonstrating high success rates in real-world experiments.

Versatile Variable-Stiffness Scooping End-Effector: Tilting-Scooping-Transfer Mechanism for Objects with Various Properties

Yuta Takahashi, Satoshi Tadokoro

Robotic Intelligence

🎯 What it does: Developed a variable stiffness tilting-scooping-transmission mechanism end-effector capable of uniformly scooping, grasping, and transporting various objects with different properties, and verified its feasibility through a prototype.

Vertebrae-based Global X-ray to CT Registration for Thoracic Surgeries

Lilu Liu, Yue Wang

Pose EstimationOptimizationConvolutional Neural NetworkMultimodalityBiomedical DataComputed Tomography

🎯 What it does: Proposes a global X-ray to CT registration method based on vertebrae, automatically locating vertebrae centers via CNN, introducing a 4-DoF solver and AE2 estimator, ultimately achieving an end-to-end trained registration framework under clinical settings.

VIHE: Virtual In-Hand Eye Transformer for 3D Robotic Manipulation

Weiyao Wang, Liangjun Zhang

Robotic IntelligenceTransformerVision-Language-Action ModelImage

🎯 What it does: Propose a Virtual Hand-Inside View Transformer (VIHE) that enhances the operational capabilities of 3D robotic arms through action-aware view rendering and multi-stage autoregressive action optimization.

Vine Robots that Evert through Bending

Rui Wu, Stefano Mintchev

Robotic Intelligence

🎯 What it does: Designed and implemented a vine robot that utilizes bending-induced overall rotation through multiple interconnected tubes, eliminating the need for the tip device and pressurized base required by traditional vine robots, achieving non-bending retraction.

Vinymap: a Vineyard Inspection and 3D Reconstruction Framework for Agricultural Robots

Ioannis Zarras (National Technical University Of Athens), E. Papadopoulos (National Technical University Of Athens)

Depth EstimationAutonomous DrivingRobotic IntelligenceImageAgriculture Related

🎯 What it does: Developed a framework for vineyard inspection and 3D reconstruction based on a robot platform

Virtual model control for compliant reaching under uncertainties

Yi Zhang, F. Forni

Robotic IntelligenceWorld Model

🎯 What it does: Using the VMC framework to design six virtual models, achieving reachable tasks for force-controlled robotic arms in environments with obstacles and uncertainties, and validating them on real robots.

VIRUS-NeRF - Vision, InfraRed and UltraSonic based Neural Radiance Fields

Nicolaj Schmid (ETH Zurich), Florian Tschopp (ETH Zurich)

Autonomous DrivingComputational EfficiencyNeural Radiance FieldSimultaneous Localization and MappingMultimodalityUltrasound

🎯 What it does: Proposed a VIRUS-NeRF model that utilizes visual, infrared, and ultrasonic sensors for low-cost local map construction.

ViSaRL: Visual Reinforcement Learning Guided by Human Saliency

Anthony Liang, Erdem Biyik

Convolutional Neural NetworkTransformerReinforcement LearningImage

🎯 What it does: Proposes a visual saliency guided reinforcement learning method, ViSaRL, which leverages saliency information to learn visual representations, enhancing the success rate, sample efficiency, and generalization ability of RL agents across multiple tasks from pixel inputs.

Vision-Language Model-based Physical Reasoning for Robot Liquid Perception

Wen Lai, T. Lam

Robotic IntelligenceLarge Language ModelVision Language ModelImageTime Series

🎯 What it does: Propose a paradigm leveraging GPT-4V to enable robots to perceive liquid objects through visualized environmental feedback, and use visualized non-visual feedback (e.g., time series diagrams of F/T sensor data) for physical understanding.

Visual Attention Based Cognitive Human–Robot Collaboration for Pedicle Screw Placement in Robot-Assisted Orthopedic Surgery

Chen Chen, Xiang Li

Robotic IntelligenceTransformerBiomedical Data

🎯 What it does: Propose a cognitive human-robot collaboration framework for robot-assisted spinal pedicle screw implantation surgery, which includes an intuitive AR-haptic human-robot interface, a surgeon model based on visual attention, and a shared interactive control scheme.

Visual Forecasting as a Mid-level Representation for Avoidance

Hsuan-Kung Yang, Chun-Yi Lee

Autonomous DrivingRepresentation LearningWorld ModelVideo

🎯 What it does: Proposes visual prediction as an intermediate representation to improve navigation and obstacle avoidance in dynamic environments, and validates two visual prediction strategies (bounding box sequences and enhanced paths)

Visual Imitation Learning of Task-Oriented Object Grasping and Rearrangement

Yichen Cai, Tamim Asfour

Robotic IntelligenceReinforcement LearningNeural Radiance FieldVideo

🎯 What it does: Propose a multi-feature implicit model called MIMO, and design a framework for learning task-oriented grasping and rearrangement from single/multi-view demonstration videos based on this model.

Visual Loop Closure Detection with Thorough Temporal and Spatial Context Exploitation

Jiaxin Li, Wei Liang

Convolutional Neural NetworkRecurrent Neural NetworkSimultaneous Localization and MappingImageVideo

🎯 What it does: Proposes an algorithm called TOSA for efficient visual loop closure detection that leverages temporal and spatial context.

Visual Perception System for Autonomous Driving

Qi Zhang, Wenbin Li

Object DetectionObject TrackingAutonomous DrivingSimultaneous Localization and Mapping

🎯 What it does: This study proposes a vision-based autonomous driving perception system that combines trajectory tracking and prediction of moving objects with environmental localization and mapping to achieve collision prevention and navigation support.

Visual Place Recognition in Unstructured Driving Environments

Utkarsh Rai (International Institute Of Information Technology), C. V. Jawahar (International Institute Of Information Technology)

RetrievalAutonomous DrivingImageBenchmark

🎯 What it does: Propose an Indian driving visual position recognition (VPR) dataset and develop an interactive image-to-image annotation tool, evaluating and providing quantitative and qualitative analysis of multiple matching methods.

Visual Preference Inference: An Image Sequence-Based Preference Reasoning in Tabletop Object Manipulation

Joonhyung Lee, Sungjoon Choi

Robotic IntelligenceReinforcement Learning from Human FeedbackPrompt EngineeringVision-Language-Action ModelImageSequentialChain-of-Thought

🎯 What it does: In a desktop object manipulation environment, human preferences are inferred from a series of raw visual observations, and the Chain-of-Visual-Residuals (CoVR) method is proposed for visual reasoning.

Visual Quality Inspection Planning: A Model-Based Framework for Generating Optimal and Feasible Inspection Poses

Vanessa Staderini, Andreas Kugi

OptimizationRobotic Intelligence

🎯 What it does: Propose a model-based visual quality inspection planning framework that generates optimal inspection poses tailored to object geometry and robot kinematics;

Visual Servo Control of a Conceptual Magnetically Anchored and Guided Flexible Endoscope

Weibing Li, Yongping Pan

OptimizationRobotic Intelligence

🎯 What it does: This paper designs a magnetic fixation and guidance flexible endoscope, and implements the function of autonomous tracking of surgical instruments based on vision servo control.

Visual Timing For Sound Source Depth Estimation in the Wild

Wei Sun, Lili Qiu

Depth EstimationOptical FlowMultimodality

🎯 What it does: Propose the FBDepth framework, which utilizes photoacoustic time-of-flight (ToF) for passive audio-visual depth estimation, integrates semantic features with physical space cues, designs a coarse-to-fine pipeline to improve localization accuracy, and employs visual timestamps, audio clips, and object visual features for source depth regression.

Visual-Geometry GP-based Navigable Space for Autonomous Navigation

Mahmoud Ali, Lantao Liu

Autonomous Driving

🎯 What it does: Propose a spatial modeling framework called VG-SGP that simultaneously considers semantic and geometric information for autonomous navigation in unknown environments.

Visuo-Tactile Exploration of Unknown Rigid 3D Curvatures by Vision-Augmented Unified Force-Impedance Control

Kübra Karacan, Sami Haddadin

Robotic IntelligenceMultimodality

🎯 What it does: Propose a vision-enhanced unified force-damping control (VA-UFIC) for intuitive detection of unknown 3D curvature through collaborative vision and tactile sensing.

Visuo-Tactile Zero-Shot Object Recognition with Vision-Language Model

Shiori Ueda, Hideo Saito

RecognitionVision Language ModelTextMultimodality

🎯 What it does: This paper proposes a method to integrate tactile data into vision-language models for visual-tactile zero-shot object recognition.

VIVO: A Visual-Inertial-Velocity Odometry with Online Calibration in Challenging Condition

Fuzhang Han, Rong Xiong

OptimizationRobotic IntelligenceSimultaneous Localization and MappingMultimodality

🎯 What it does: Proposes a VIVO framework that tightly couples speed measurements with MSCKF-based visual inertial odometry and performs online calibration of external parameters, applicable to both wheeled and legged robots;

VLPG-Nav: Object Navigation Using Visual Language Pose Graph and Object Localization Probability Maps

Senthil Hariharan Arul, Dinesh Manocha

Pose EstimationRobotic IntelligenceGraph Neural NetworkVision Language ModelMultimodality

🎯 What it does: Proposed a visual language navigation method called VLPG-Nav, aiming to guide robots to locate specified objects in home environments and center the objects within the camera's field of view.

Voltage Regulation in Polymer Electrolyte Fuel Cell Systems Using Gaussian Process Model Predictive Control

Xiufei Li, Miao Yang

OptimizationPhysics Related

🎯 What it does: Propose a method using a Gaussian process model predictive control (MPC) to stabilize the output voltage of polymer electrolyte fuel cells (PEFC) by adjusting hydrogen and air flow rates.

Volumetric Mapping with Panoptic Refinement using Kernel Density Estimation for Mobile Robots

K. Nguyen, M. Huber

Depth EstimationRobotic IntelligenceConvolutional Neural Network

🎯 What it does: Refine panoramic segmentation of 3D scenes using kernel density estimation, remove anomalies in depth perception, and reconstruct using projected signed distance functions (SDF).

Volumetric Semantically Consistent 3D Panoptic Mapping

Yang Miao, Dániel Baráth

Autonomous DrivingOptimizationSimultaneous Localization and Mapping

🎯 What it does: Propose an online 2D-to-3D semantic instance mapping algorithm for generating comprehensive, accurate, and efficient semantic 3D maps, applicable to unmanned systems operating in unstructured environments.

VoxelContrast: Voxel Contrast-Based Unsupervised Learning for 3D Point Clouds

Yuxiang Qin, Hao Sun

ClassificationObject DetectionAutonomous DrivingContrastive LearningPoint Cloud

🎯 What it does: Propose an unsupervised learning method called VoxelContrast based on voxel contrast, which uses voxelization to preprocess point clouds and integrates voxel information into contrastive learning, utilizing instance discrimination as a proxy task for model pre-training.

VRExplorer: An Efficient View-Region based Autonomous Exploration Method in Unknown Environments for UAV

Kai Xu, Hui Cheng

OptimizationRobotic IntelligenceSimultaneous Localization and Mapping

🎯 What it does: Proposes a UAV autonomous exploration method based on the concept of view regions, replacing traditional view point generation and selection processes to simplify exploration planning;

VRSO: Visual-Centric Reconstruction for Static Object Annotation

Chenyao Yu, Cong Yang

Depth EstimationAutonomous DrivingImage

🎯 What it does: Proposed a vision-based static object annotation method called VRSO, which automatically recovers static objects in 3D space and generates annotations using only camera images.

WasteGAN: Data Augmentation for Robotic Waste Sorting through Generative Adversarial Networks

Alberto Bacchin, Takuya Kiyokawa

SegmentationData SynthesisRobotic IntelligenceGenerative Adversarial NetworkImage

🎯 What it does: Proposed the WasteGAN data augmentation method to enhance the semantic segmentation performance of robot garbage classification under scenarios with extremely limited labeled samples (e.g., 100 samples), and utilized the enhanced segmentation results to achieve semantic-aware grasping poses, improving garbage identification and separation efficiency.

Waypoint-Based Reinforcement Learning for Robot Manipulation Tasks

Shaunak A. Mehta, Dylan P. Losey

Robotic IntelligenceReinforcement Learning

🎯 What it does: Propose a reinforcement learning method for robot manipulation based on waypoints, where the robot learns waypoint trajectories and then uses existing controllers for interpolation.

Weakly Scene Segmentation Using Efficient Transformer

Hao Huang, Yi Fang

SegmentationTransformerPoint Cloud

🎯 What it does: Proposes a weakly supervised large-scale indoor point cloud scene semantic segmentation method, requiring only 1‰ of points to be labeled, and develops an efficient point neighborhood Transformer and low-rank sparse self-attention approximation.

Wheelchair Maneuvering with a Single-Spherical-Wheeled Balancing Mobile Manipulator

Cunxi Dai, Ralph Hollis

Robotic Intelligence

🎯 What it does: Propose a whole-body motion planning control framework based on quasi-static analysis to achieve dynamic stable control of heavy non-homogeneous wheelchairs on a single-ball-wheel balancing mobile manipulator.

When, What, and with Whom to Communicate: Enhancing RL-based Multi-Robot Navigation through Selective Communication

Senthil Hariharan Arul, Dinesh Manocha

Robotic IntelligenceTransformerReinforcement Learning

🎯 What it does: Propose an RL-based decentralized navigation method that learns when, what information, and with which neighbors to communicate to achieve safe collaborative navigation.

Where and When Should the Teleoperated Avatar Look: Gaze Instruction Dataset for Enhanced Teleoperated Avatar Communication*

Kenya Hoshimure, Hiroshi Ishiguro

Robotic Intelligence

🎯 What it does: Collected and annotated a gaze position dataset considering the context of avatars and operators, followed by dynamic area of interest (AOI) analysis to explore gaze proportions under different contexts.

Whleaper: A 10-DOF Flexible Bipedal Wheeled Robot

Yinglei Zhu, Jianyu Chen

Robotic Intelligence

🎯 What it does: Proposed and implemented a 10-degree-of-freedom (10-DOF) bipedal wheeled-leg robot named Whleaper, introducing a 3-DOF design at the hip of each leg, enhancing its stability and flexibility.

Whole-body Compliance Control for Quadruped Manipulator with Actuation Saturation of Joint Torque and Ground Friction

Tianlin Zhang, Yunjiang Lou

OptimizationRobotic Intelligence

🎯 What it does: Proposed a whole-body compliance controller that combines a set-point feedback impedance scheme with hierarchical quadratic programming (HQP) to prevent unstable behaviors in quadruped manipulators under actuator saturation conditions.

Whole-body Humanoid Robot Locomotion with Human Reference

Qiang Zhang, Renjing Xu

Robotic Intelligence

🎯 What it does: Developed a full-scale humanoid robot Adam and proposed an imitation learning framework based on adversarial motion priors, achieving a human-like gait

WidthFormer: Toward Efficient Transformer-based BEV View Transformation

Chenhongyi Yang, Elliot Crowley

Autonomous DrivingTransformerImageBenchmark

🎯 What it does: Proposed WidthFormer, a Transformer-based module for generating real-time and efficient bird's-eye view (BEV) representations from multi-view cameras, applicable to autonomous driving scenarios.

Wing twist and folding work in synergy to propel flapping wing animals and robots

Xiaozhou Fan, Kenneth Breuer

Robotic IntelligenceOptical FlowPhysics Related

🎯 What it does: Designed and built a three-degree-of-freedom flapping robot named Flapperoo to study the aerodynamic benefits of wing folding and twisting.

Wirelessly Actuated Rotation-free Magnetic Motor

Umur Ulas Harman, Shuhei Miyashita

Robotic IntelligencePhysics Related

🎯 What it does: Developed a millimeter-scale non-rotating motor wirelessly driven by an external magnetic field;

Working Backwards: Learning to Place by Picking

Oliver Limoyo, Gregory Dudek

Robotic IntelligenceImage

🎯 What it does: Automatically collects placement task demonstrations through reverse grasping and trains a behavioral cloning policy based on visual observations, enabling robots to complete object placement tasks in unconstrained environments.

WSCLoc: Weakly-Supervised Sparse-View Camera Relocalization via Radiance Field

Jialu Wang, Niki Trigoni

Pose EstimationNeural Radiance FieldMultimodality

🎯 What it does: Propose the WSCLoc system, which adopts a two-stage weakly supervised sparse perspective camera relocalization method, utilizing WFT-NeRF and WFT-Pose for joint optimization, and enhancing learning stability through temporal information and time-encoding based random perspective synthesis.

X-neuron: Interpreting, Locating and Editing of Neurons in Reinforcement Learning Policy

Yuhong Ge, Dahua Lin

Explainability and InterpretabilityReinforcement Learning

🎯 What it does: Propose a three-step process—interpreting neurons, locating X-neurons, and editing their activation values—to enhance the interpretability and human controllability of reinforcement learning strategies.

X-Ray-Guided Magnetic Fields for Wireless Control of Untethered Magnetic Robots in Cerebral Vascular Phantoms

Leendert-Jan W. Ligtenberg, Islam S. M. Khalil

Robotic IntelligenceBiomedical DataComputed Tomography

🎯 What it does: Studied the wireless control of untangled magnetic robots (UMR) in cerebral vascular models using X-ray guided magnetic fields to enhance the precision and maneuverability of neurosurgical procedures.

Zero-Shot Transfer of a Tactile-based Continuous Force Control Policy from Simulation to Robot

Luca Lach, Carme Torras

Domain AdaptationRobotic IntelligenceReinforcement Learning

🎯 What it does: Propose a model-free, tactile-driven continuous force control strategy based on deep reinforcement learning, trained in simulation and directly transferred to robots without further fine-tuning.

Zero123-6D: Zero-shot Novel View Synthesis for RGB Category-level 6D Pose Estimation

F. D. Felice, C. Avizzano

Pose EstimationDiffusion modelImage

🎯 What it does: Propose the Zero123-6D method, which combines a zero-shot novel view synthesizer based on diffusion models with feature extraction techniques for category-level 6D pose estimation of RGB images, and refines the coarse pose through an online optimization method.

ν-DBA: Neural Implicit Dense Bundle Adjustment Enables Image-Only Driving Scene Reconstruction

Yunxuan Mao, Yue Wang

Autonomous DrivingOptimizationSupervised Fine-TuningSimultaneous Localization and MappingOptical FlowImage

🎯 What it does: Proposed the ν-DBA framework, which utilizes three-dimensional neural implicit surfaces for dense bundle adjustment, jointly optimizes sensor trajectory and 3D map, guides optimization through geometric errors predicted by dense optical flow, and performs scene-specific self-supervised fine-tuning of the optical flow model to enhance dense reconstruction quality.