IROS 2025 Papers — Page 9
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1984 papers
Generalizable Humanoid Manipulation with 3D Diffusion Policies
Yanjie Ze, Jiajun Wu
Robotic IntelligenceDiffusion modelPoint Cloud
🎯 What it does: A real-world robotic system was constructed, integrating full-body teleoperation data acquisition, a 25-degree-of-freedom (DoF) full-scale humanoid robot platform (with adjustable-height tracks and 3D LiDAR), and an improved 3D diffusion policy learning algorithm, enabling the humanoid robot to autonomously complete manipulation tasks in diverse real-world scenarios using data from a single scene and onboard computation.
Generalizable Image Repair for Robust Visual Control
Carson Sobolewski, Ivan Ruchkin
Image TranslationRestorationGenerative Adversarial NetworkImage
🎯 What it does: Proposed a real-time image restoration module that repairs damaged images before the controller is used, utilizing CycleGAN for unpaired image translation, pix2pix to improve quality with paired data, and introducing a perceptual consistency loss tailored for control performance.
Generalized Locomotion in Out-of-distribution Conditions with Robust Transformer
Lingxiao Guo, Yue Gao
Robotic IntelligenceTransformer
🎯 What it does: Investigated the robustness of legged robots walking in dynamic and perceptual noise environments outside of training, and proposed a new Transformer variant called ROLT
Generating Actionable Robot Knowledge Bases by Combining 3D Scene Graphs with Robot Ontologies
G. Nguyen, Michael Beetz
Robotic IntelligenceGraph
🎯 What it does: Developed a unified scene graph model that standardizes multiple scene description formats (MJCF, URDF, SDF) into USD format, and integrates the scene graph with the robot's body through semantic reports, achieving the transformation of complex environmental data into executable knowledge; performs USD conversion, semantic annotation, and knowledge graphification for procedural 3D environments to answer capability questions, and developed a web-based visualization tool to support semantic mapping;
GeoFlow-SLAM: A Robust Tightly-Coupled RGBD-Inertial and Legged Odometry Fusion SLAM for Dynamic Legged Robotics
Tingyang Xiao, Zhizhong Su
OptimizationRobotic IntelligenceSimultaneous Localization and MappingOptical FlowImageMultimodality
🎯 What it does: Propose GeoFlow-SLAM, which integrates RGB-D, IMU, and leg odometry for quadruped robots with rapid motion.
Geometric Retargeting: A Principled, Ultrafast Neural Hand Retargeting Algorithm
Zhao-Heng Yin, Mustafa Mukadam
Pose EstimationRobotic Intelligence
🎯 What it does: Proposed a ultra-high-speed neural hand reorientation algorithm called GeoRT, which can convert human finger keypoints into robot hand keypoints at a frequency of 1KHz.
GeoSafe: A Unified Unconstrained Multi-DOF Optimization Framework for Multi-UAV Cooperative Hoisting and Obstacle Avoidance
Xingyu Li, Sen Mei
Optimization
🎯 What it does: Proposes an optimization-based framework that incorporates rotational degrees of freedom to achieve multi-UAV collaborative lifting and obstacle avoidance. It transforms constrained formation adjustments into unconstrained optimization via MINCO transformation and penalty functions, further enhancing safe passage through iterative region expansion and semidefinite programming in the GeoSafe algorithm.
GeoScene: Temporal 3D Semantic Scene Completion with Geometric Correlation between Images
Xiaoyu Zhu, Kangcheng Liu
SegmentationDepth EstimationAutonomous DrivingImagePoint Cloud
🎯 What it does: Proposes a multi-frame matching framework called GeoScene, which reconstructs spatial structures and infers scene semantics by leveraging geometric correlations between sequential images.
GeT-USE: Learning Generalized Tool Usage for Bimanual Mobile Manipulation via Simulated Embodiment Extensions
Bohan Wu, Roberto Martín-Martín
Robotic IntelligenceWorld ModelImage
🎯 What it does: Exploring the expansion of robotic physical capabilities in simulation, learning to recognize general tool geometry, and transferring this knowledge to a real 22-degree-of-freedom dual-arm mobile robot for vision-driven tool use
GFM-Planner: Perception-Aware Trajectory Planning with Geometric Feature Metric
Yue Lin, Huchuan Lu
Autonomous DrivingRobotic IntelligenceSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposes a perception-aware trajectory planning framework called GFM-Planner based on geometric feature metrics, which utilizes LiDAR localization accuracy metrics to guide the robot to avoid low-feature regions
GHO-WBC: A Gradient-Based Hierarchical Kinematic Optimization Approach to Enhance the Reachability of a Humanoid Robot
Weiliang Zhu, Yibin Li
OptimizationRobotic Intelligence
🎯 What it does: Proposed a gradient-based whole-body kinematics optimization method called GHO-WBC to enhance the reachable workspace of the end-effector of humanoid robots when maintaining a stationary posture.
GIANT - Global Path Integration and Attentive Graph Networks for Multi-Agent Trajectory Planning
Jonas le Fevre Sejersen, Erdal Kayacan
Autonomous DrivingRobotic IntelligenceGraph Neural Network
🎯 What it does: Propose a multi-robot collision avoidance method that integrates global path planning with local navigation, using an attention graph neural network to handle dynamic interactions.
GIPD: Global Intent Prediction and Decomposition of Cooperative Multi-Robot System in Non-Communication Environments
Yu Zhao, Haibin Shao
Robotic IntelligenceBenchmark
🎯 What it does: Proposes the Global Intent Prediction and Decomposition (GIPD) framework, enabling robots to autonomously infer globally consistent intentions, assign responsibilities, and select appropriate tasks through local observations without relying on communication, thereby achieving implicit collaboration and effective execution.
GLIC-Calib: Targetless and One-Shot Spatial-Temporal Calibration of LiDAR-IMU-Camera for Ground Vehicles
Hang Zhao, Dongyan Wei
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Developed a target-agnostic, one-time spatiotemporal calibration method for LiDAR-IMU-camera systems
Global Optimization of Stochastic Black-Box Functions with Arbitrary Noise Distributions using Wilson Score Kernel Density Estimation
Thorbjørn Mosekjær Iversen, Henrik Gordon Petersen
Optimization
🎯 What it does: Provide confidence bounds for stochastic black-box functions with arbitrary noise distributions using the Wilson Score Kernel Density Estimator (WS-KDE), achieving global optimization in simulation experiments and an automatic trap design task for vibrating part feeders.
GO-Flock: Goal-Oriented Flocking in 3D Unknown Environments with Depth Maps
Yan Rui Tan, R. Teo
Depth EstimationRobotic IntelligenceImage
🎯 What it does: Developed a GO-Flock hybrid swarm flying framework that integrates planning with reactive artificial potential field control, using depth graph extraction of waypoints and virtual agents to achieve UAV swarm navigation in 3D unknown environments.
Go-SLAM: Grounded Object Segmentation and Localization with Gaussian Splatting SLAM
Phu-Cuong Pham, Aniket Bera
Object DetectionSegmentationGaussian SplattingSimultaneous Localization and MappingImage
🎯 What it does: Integrate 3D Gaussian Splatting SLAM with object-based segmentation and open vocabulary queries to achieve object-aware 3D scene reconstruction, incrementally building high-precision 3D maps under RGB-D input, while assigning unique object IDs to Gaussian primitives to embed semantic information.
GO-VMP: Global Optimization for View Motion Planning in Fruit Mapping
A. Jose, Maren Bennewitz
OptimizationRobotic IntelligenceAgriculture Related
🎯 What it does: Proposed a globally optimized viewpoint motion planning method aimed at maximizing fruit coverage while minimizing robot motion cost.
GOEN: Guided Obstacle Endpoint Navigation for Real-Time Collision-Free Path Planning in Unstructured Environments
Zhicheng Zhao, Yibin Li
Robotic IntelligencePoint Cloud
🎯 What it does: Designed and implemented the GOEN navigation and path planning framework, which processes 3D point clouds in real-time, generates and iteratively optimizes intermediate waypoints, and outputs kinematically feasible and collision-free trajectories.
GPGS: Geometric Priors for 3D Gaussian Splatting in Structural Environments
Ziwei Xu, Shunbo Zhou
Gaussian Splatting
🎯 What it does: Propose the GPGS method, which directly constrains Gaussian Splatting in 3D space using geometric priors to enhance geometric representation in structured environments.
GRaD-Nav: Efficiently Learning Visual Drone Navigation with Gaussian Radiance Fields and Differentiable Dynamics
Qianzhong Chen, Mac Schwager
Autonomous DrivingReinforcement LearningNeural Radiance FieldGaussian SplattingImage
🎯 What it does: Proposes a visual drone navigation framework GRaD-Nav that combines 3D Gaussian point mist technology with differentiable depth reinforcement learning, and verifies its training efficiency and zero-shot sim-to-real transfer capability in hardware experiments.
Gradient Field-Based Dynamic Window Approach for Collision Avoidance in Complex Environments
Ze Zhang, Knut Åkesson
Autonomous DrivingSafty and PrivacyRobotic Intelligence
🎯 What it does: Proposed an improved gradient field-based dynamic window approach (GF-DWA) for safe and flexible navigation of multi-robot systems in complex environments;
Graph2Scene: Versatile 3D Indoor Scene Generation with Interaction-aware Scene Graph
Minglin Chen, Yulan Guo
GenerationGraph Neural NetworkTransformerLarge Language ModelGraph
🎯 What it does: Propose a 3D indoor scene generation method based on interactive perception scene graphs, achieving a balance between high diversity and fine-grained control.
GraphGarment: Learning Garment Dynamics for Bimanual Cloth Manipulation Tasks
Wei Chen, Petar Kormushev
Robotic IntelligenceGraph Neural NetworkGraph
🎯 What it does: Propose the GraphGarment method, which learns a clothing dynamics model based on robot control inputs through graph neural networks, and applies it to achieve clothing manipulation tasks such as two-handed grasping and hanging.
GraphSCENE: On-Demand Critical Scenario Generation for Autonomous Vehicles in Simulation
Efimia Panagiotaki, Daniele De Martini
Autonomous DrivingGraph Neural Network
🎯 What it does: Propose a dynamic temporal scene graph generation method based on graph neural networks, which can generate safety-critical traffic scenarios in simulation environments according to user-specified AV behaviors, dynamic agent sets, and criticality levels.
Grasp EveryThing (GET): 1-DoF, 3-Fingered Gripper with Tactile Sensing for Robust Grasping
Michael Burgess, Edward H. Adelson
Robotic IntelligenceImage
🎯 What it does: Proposed a 1-DoF, 3-finger gripper named Grasp EveryThing (GET), capable of adaptively grasping objects with different geometries and achieving stable grasping.
Grasping and Alignment of Stacked Fabrics by Robot Hands with Sticky Fingers
K. Kondo, Shunsuke Kudoh
Pose EstimationRobotic Intelligence
🎯 What it does: Proposed and implemented a dual-arm robot method that only grasps the top layer of stacked fabric and places it at a target location.
GRASPLAT: Enabling dexterous grasping through novel view synthesis
Matteo Bortolon (ETH Zurich), A. D. Bue (ETH Zurich)
Pose EstimationRobotic IntelligenceGaussian SplattingImage
🎯 What it does: Propose the GRASPLAT framework, which uses only RGB images to regress hand joints by synthesizing physically plausible hand-object interaction images, achieving dexterous grasping.
GraspMamba: A Mamba-based Language-driven Grasp Detection Framework with Hierarchical Feature Learning
H. Nguyen, M. Vu
Object DetectionRobotic IntelligenceVision-Language-Action ModelMultimodality
🎯 What it does: Proposed the GraspMamba framework for language-driven grasping detection, combining the Mamba visual backbone with hierarchical feature fusion;
GraspMAS: Zero-Shot Language-driven Grasp Detection with Multi-Agent System
Quang Nguyen, Anh Nguyen
Robotic IntelligenceAgentic AIVision-Language-Action Model
🎯 What it does: Proposed and implemented a multi-agent system called GraspMAS for zero-shot language-driven grasping detection.
Gravity Compensation with Dual Quaternions
F. J. A. García, Yuichi Kobayashi
Robotic IntelligencePhysics Related
🎯 What it does: Propose a centroid representation based on dual quaternions for efficiently computing the gravity compensation term on rigid robots with arbitrary kinematic chain structures and arbitrary base coordinate systems.
GRIP: A General Robotic Incremental Potential Contact Simulation Dataset for Unified Deformable-Rigid Coupled Grasping
Siyu Ma, Chenfanfu Jiang
Robotic IntelligenceBenchmark
🎯 What it does: Constructed and released the GRIP dataset, which automatically generated 1200 objects and 100,000 grasping poses of soft and hard grasping interaction data through Incremental Potential Contact (IPC) simulation technology.
Growing manipulators through feeding material outside-in: Inversion robots
Xinyi Pi, Helge A. Wurdemann
Robotic Intelligence
🎯 What it does: This paper proposes a growth-type robot called the reversal robot, based on an outward-to-inward material feeding mechanism, and investigates its design, implementation, and experimental validation. It explores growth behavior under different pressures and diameters, navigation capability along a predetermined trajectory, and the carrying effectiveness of the end-effector tool.
GS-SDF: LiDAR-Augmented Gaussian Splatting and Neural SDF for Geometrically Consistent Rendering and Reconstruction
Jianheng Liu, Fu Zhang
Gaussian SplattingPoint Cloud
🎯 What it does: Proposes a unified LiDAR-visual system that combines efficient Gaussian splatting with neural signed distance fields (SDF) to achieve geometrically consistent rendering and reconstruction.
GSO-SLAM: Robust Monocular SLAM with Global Structure Optimization
Bingzheng Jiang, Han Ding
OptimizationSimultaneous Localization and MappingImage
🎯 What it does: Propose a robust monocular visual SLAM system that combines point, line, and vanishing point features to achieve accurate camera pose estimation and map construction.
GSplatLoc: Grounding Keypoint Descriptors into 3D Gaussian Splatting for Improved Visual Localization
G. Sidorov, S. Kolyubin
Pose EstimationGaussian SplattingSimultaneous Localization and Mapping
🎯 What it does: Developed a visual localization framework GSplatLoc that combines 3D Gaussian Splatting technology with structured keypoint matching;
GSplatVNM: Point-of-View Synthesis for Visual Navigation Models Using Gaussian Splatting
Kohei Honda, Ryo Yonetani
Data SynthesisAutonomous DrivingGaussian SplattingImage
🎯 What it does: Combine 3D Gaussian Splatting (3DGS) with a Visual Navigation Model (VNM) to propose the GSplatVNM framework, which synthesizes intermediate viewpoints to fill gaps in sparse image databases and reduce storage costs.
GSPR: Multimodal Place Recognition Using 3D Gaussian Splatting for Autonomous Driving
Zhangshuo Qi, Guangming Xiong
RetrievalAutonomous DrivingGraph Neural NetworkTransformerGaussian SplattingImageMultimodalityPoint Cloud
🎯 What it does: Proposed a multimodal localization network called GSPR based on 3D Gaussian mapping, which fuses multi-view RGB images with LiDAR point clouds into a unified spatiotemporal scene representation and extracts global descriptors through 3D graph convolution and Transformer.
GTAD: Global Temporal Aggregation Denoising Learning for 3D Semantic Occupancy Prediction
Tianhao Li, Weifeng Ge
SegmentationAutonomous DrivingTransformerAuto EncoderVideoPoint CloudBenchmark
🎯 What it does: Proposed a global temporal aggregation denoising network called GTAD for 3D semantic occupancy prediction, which can simultaneously utilize local temporal features of the current frame and global temporal information from the historical sequence;
HAC-LOCO: Learning Hierarchical Active Compliance Control for Quadruped Locomotion under Continuous External Disturbances
Xiang Zhou, Qingrui Zhang
Robotic IntelligenceReinforcement LearningAuto Encoder
🎯 What it does: Proposed and implemented a two-stage hierarchical learning framework for quadruped robots to achieve active compliance control under continuous external disturbances; the first stage trains a velocity tracking policy and an auto-encoder, and builds a force estimator through supervised learning; the second stage learns a compliance action module based on pre-trained encoders and policies, using real-time force estimation to actively adjust velocity commands.
HACTS: a Human-As-Copilot Teleoperation System for Robot Learning
Zhiyuan Xu, Jian Tang
Robotic IntelligenceReinforcement Learning from Human FeedbackReinforcement LearningSequential
🎯 What it does: This paper proposes the HACTS system, achieving bidirectional real-time synchronization between the robot arm and remote control hardware, and employs a steering-wheel-like feedback mechanism to enable human co-pilots to intervene in real time, while collecting action-correction data for subsequent learning.
HAMF: A Hybrid Attention-Mamba Framework for Joint Scene Context Understanding and Future Motion Representation Learning
Xiaodong Mei, Dan Xu
Autonomous DrivingRepresentation LearningTransformerImagePoint CloudBenchmark
🎯 What it does: Proposed the HAMF framework for joint scene context understanding and future motion representation learning to improve motion prediction.
Handheld Confocal Endomicroscope System with Tremor Compensation for Retinal Imaging
Myung Ho Lee, Cheol Song
Recurrent Neural NetworkBiomedical Data
🎯 What it does: This paper proposes a new handheld confocal endoscopy system, which achieves non-contact, high-resolution retinal imaging by integrating a self-made imaging probe, an optical coherence tomography (OCT) distance sensor, and motor-assisted jitter suppression.
HannesImitation: Grasping with the Hannes Prosthetic Hand via Imitation Learning
Carlo Alessi (Italian Institute of Technology), Lorenzo Natale (Italian Institute of Technology)
Robotic IntelligenceDiffusion modelImage
🎯 What it does: Propose HannesImitationPolicy, which uses imitation learning to control the Hannes prosthetic hand for object grasping in unstructured environments, and construct the HannesImitationDataset;
HAPI: A Model for Learning Robot Facial Expressions from Human Preferences
Dongsheng Yang, Shin’ya Nishida
GenerationRobotic IntelligenceReinforcement Learning from Human FeedbackContrastive Learning
🎯 What it does: Propose a learning-ranking framework based on human feedback and develop the Human Affective Pairwise Impressions (HAPI) model to improve robot facial expression generation.
Haptic Feedback Control Strategy for Microswarm Navigation in Flowing Environments
Ying Cao, Qianqian Wang
Robotic Intelligence
🎯 What it does: Propose a real-time navigation and control strategy for magnetic microswarms based on tactile feedback, which can real-time sense and adjust the position and shape of microswarms in fluid environments;
Haptic Shared Control of a Pair of Microrobots for Telemanipulation using Constrained Optimization
Leon Raphalen, Claudio Pacchierotti
OptimizationRobotic Intelligence
🎯 What it does: Designed and implemented a tactile sharing control system for magnetic microrobot teleoperation.
Haptic-ACT: Bridging Human Intuition with Compliant Robotic Manipulation via Immersive VR
Kelin Li, Petar Kormushev
Robotic IntelligenceTransformer
🎯 What it does: Developed an immersive VR remote control platform for collecting demonstrations from remote human operators, and proposed the Haptic Action Chunking with Transformers (Haptic-ACT) framework.
Haptic-Informed ACT with a Soft Gripper and Recovery-Informed Training for Pseudo Oocyte Manipulation
P. M. U. Eljuri, Tadahiro Taniguchi
Robotic IntelligenceTransformer
🎯 What it does: Proposed a haptic-aware ACT (Action Chunking with Transformers) system for operating pseudo-oocytes, and introduced 3D-printed TPU soft grippers to achieve precise grasping
HARP-NeXt: High-Speed and Accurate Range-Point Fusion Network for 3D LiDAR Semantic Segmentation
Samir Abou Haidar, Jean-Emmanuel Deschaud
SegmentationAutonomous DrivingConvolutional Neural NetworkPoint CloudBenchmark
🎯 What it does: Proposed a high-speed and high-accuracy LiDAR semantic segmentation network called HARP-NeXt, along with an efficient preprocessing method and a multi-scale range-point fusion backbone.
HD-OOD3D: Supervised and Unsupervised Out-of-Distribution object detection in LiDAR data
Louis Soum-Fontez, Franccois Goulette
Object DetectionAnomaly DetectionHyperparameter SearchPoint Cloud
🎯 What it does: Propose the HD-OOD3D two-stage method for detecting unknown objects in LiDAR data
Head-mounted Robotic Needle Positioning: Learning from Augmented Reality Demonstration of Neuronavigation and Planning
Zhiwei Fang, Hongliang Ren
Robotic IntelligenceBiomedical Data
🎯 What it does: Developed a system that utilizes augmented reality (AR) to display surgical scenes and navigation information, achieving neurosurgery robot needle positioning through learning from demonstrations.
HEATS: A Hierarchical Framework for Efficient Autonomous Target Search with Mobile Manipulators
Hao Zhang, Haoyao Chen
OptimizationRobotic Intelligence
🎯 What it does: Proposes HEATS, a hierarchical and efficient autonomous target search framework designed for mobile manipulators.
HeightAware-BEV: Height-Aware Feature Mapping for Efficient Bird’s-Eye-View Perception
Renjie Zhou, Zhengjun Wang
Autonomous DrivingPoint Cloud
🎯 What it does: Proposed the HeightAware-BEV framework, achieving efficient and accurate 2D-3D view conversion through height-aware feature mapping.
Helpful DoggyBot: Open-World Object Fetching using Legged Robots and Vision-Language Models
Qi Wu, Chelsea Finn
Robotic IntelligenceVision Language ModelVision-Language-Action ModelImageTextMultimodality
🎯 What it does: Proposed a quadrupedal mobile manipulation system for indoor environments that can complete grasping tasks without real-world training.
HeStIa: Asynchronous Embodied Dynamic Locomotion Learning for Walking Robots through Multimodal Large Language Models
X. Tan, Xihe Qiu
Robotic IntelligenceTransformerLarge Language ModelVision-Language-Action ModelMultimodality
🎯 What it does: Proposed the HeStIa framework, which leverages multimodal large language models to associate robotic visual observations with motion control, enabling robots to adjust locomotion through dual-modal understanding of vision and language.
Heterogeneous Graph Network-Based UWB Localization for Complex Indoor Environments
Bo Yang, Hong Zhang
Robotic IntelligenceGraph Neural NetworkGraph
🎯 What it does: Propose a UWB localization method based on heterogeneous graph networks for mobile robots in complex indoor environments
Heterogeneous Mixed Traffic Control and Coordination
Iftekharul Islam, K. Heaslip
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: Studying the control and coordination of heterogeneous mixed traffic with robot vehicles (RVs) at urban non-signalized intersections, and evaluating their impact on traffic flow.
Heterogeneous Multi-Agent Learning in Isaac Lab: Scalable Simulation for Robotic Collaboration
Jacob Haight, Mario Harper
Robotic IntelligenceReinforcement Learning
🎯 What it does: Extended and implemented a high-fidelity physics simulation framework for heterogeneous multi-agent learning in Isaac Lab, including diverse MARL environments, HARL algorithm integration, and a scalable GPU-accelerated training system.
Heteroscedastic Bayesian Optimization-Based Dynamic PID Tuning for Accurate and Robust UAV Trajectory Tracking
Fuqiang Gu, Huidong Liu
OptimizationRobotic Intelligence
🎯 What it does: Proposed an HBO-PID algorithm that combines heteroscedastic Bayesian optimization with a classical PID controller for UAV trajectory tracking.
HFDNet: High-Frequency Divergence Attention Network for Underwater Segmentation
Hongbo Xie, Chunlei Wang
SegmentationTransformerImage
🎯 What it does: Proposed a high-frequency difference attention network (HFDNet) based on Transformer, improving underwater image semantic segmentation through frequency domain analysis
HFSENet: Hierarchical Fusion Semantic Enhancement Network for RGB-T Semantic Segmentation in Annealing Furnace Operation Area
Haoyu Yuan, Tianwei Niu
SegmentationConvolutional Neural NetworkMultimodality
🎯 What it does: Proposed an HFSENet hierarchical fusion semantic enhancement network for RGB-T semantic segmentation in welding furnace work areas, which was successfully tested on an unmanned inspection vehicle.
HGDiffuser: Efficient Task-Oriented Grasp Generation via Human-Guided Grasp Diffusion Models
Dehao Huang, Hong Zhang
Pose EstimationRobotic IntelligenceTransformerDiffusion model
🎯 What it does: Proposes HGDiffuser, a task-oriented grasping generation framework based on diffusion models, which can directly generate 6-DoF task-oriented grasps in a single stage and integrate constraints from human grasp demonstrations into the sampling process.
Hierarchical Collision-Free Configuration Planning for a Soft Manipulator
Yi Shen, Zhe Liu
OptimizationRobotic Intelligence
🎯 What it does: Proposes a hierarchical collision-agnostic configuration planning framework for soft manipulators, coordinating behavior planning, configuration planning, and shape/position control to achieve obstacle avoidance and target grasping.
Hierarchical Decision-Making for Autonomous Navigation: Integrating Deep Reinforcement Learning and Fuzzy Logic in Four-Wheel Independent Steering and Driving Systems
Yizhi Wang, Peng Chen
Autonomous DrivingReinforcement Learning
🎯 What it does: A hierarchical decision-making framework for four-wheel independent steering and driving (4WISD) systems is proposed to achieve autonomous navigation.
Hierarchical Framework for Constrained Dual-Arm Cooperative Manipulation with Whole-Body Collision Avoidance
Silong Zhang, Jianmin Ji
Data SynthesisOptimizationRobotic Intelligence
🎯 What it does: Propose a hierarchical framework that combines learning-based planning with classical control theory to achieve whole-body collision avoidance for dual-arm collaborative tasks while maintaining closed-chain constraints; simultaneously propose an efficient, cost-free data generation method.
Hierarchical Question-Answering for Driving Scene Understanding Using Vision-Language Models
Safaa Abdullahi Moallim Mohamud (Korea University), Dong Seog Han (Korea University)
Autonomous DrivingComputational EfficiencySupervised Fine-TuningPrompt EngineeringVision Language ModelImageChain-of-Thought
🎯 What it does: Proposed a hierarchical question-answering (QA) method for autonomous driving scene understanding, efficiently capturing key visual elements by refining high-level and detailed sub-questions.
Hierarchical Reactive Task Planning with Temporal Logic and Visual Servoing for Bolt-Tightening Robots in Transmission Towers
Junyi You, Haibo Du
Robotic IntelligenceImage
🎯 What it does: Proposed a framework named HTP-TV, integrating hierarchical task planning, temporal logic (LTL), and visual servoing to enable adaptive task execution by an unmanned bolt-tightening robot on transmission towers.
Hierarchical Reinforcement Learning for Articulated Tool Manipulation with Multifingered Hand
Wei Xu, Xinjun Sheng
Robotic IntelligenceReinforcement LearningPoint Cloud
🎯 What it does: Propose a hierarchical goal-conditioned reinforcement learning framework that uses a multi-fingered robotic hand, with a low-level policy adjusting tool orientation and a high-level policy controlling the robotic arm to complete object grasping.
Hierarchical Trajectory Planning Method for Piano-Playing Robot
Zirui Wang, Hong Liu
Robotic Intelligence
🎯 What it does: A hierarchical trajectory planning framework is proposed, systematically integrating obstacle avoidance and acceleration constraints to achieve path planning for a piano-playing robot.
HiFAR: Multi-Stage Curriculum Learning for High-Dynamics Humanoid Fall Recovery
Penghui Chen, Mingguo Zhao
Robotic Intelligence
🎯 What it does: Proposed and implemented the HiFAR multi-stage curriculum learning framework for achieving automatic fall recovery in high dynamic humanoid robots.
High Confidence Surgical Instrument Transparency Adopting Miniature Multi-view Endoscope System
Handing Xu, Xin-Jun Liu
RestorationSuper ResolutionBiomedical Data
🎯 What it does: By adopting a multi-view miniature endoscope system, achieve transparency of surgical instruments, using multi-view technology to fill occluded regions and employing super-resolution to restore details
High DOF Tendon-Driven Soft Hand: A Modular System for Versatile and Dexterous Manipulation
Yeonwoo Jang, Jiyun Kim
Robotic Intelligence
🎯 What it does: Designed and implemented a modular high-degree-of-freedom tendon-driven soft hand, featuring a single-finger modular design, fully integrated four-channel drive components, and precise multi-task control achieved through neural network-based trajectory planning.
High temperature sterilization resistant and enclosed three-axial force-sensing surgical instrument integrated with step-reduced FBG*
Tianliang Li, Siqi Zhu
Robotic Intelligence
🎯 What it does: Proposed a high-temperature resistant, closed-type triaxial force-sensing surgical instrument that can be integrated at the end of endoscopic surgical robots, achieving triaxial force and temperature decoupling through hydrofluoric acid etching to realize dual reflection spectroscopy.
High-dynamic Tactile Sensing for Tactile Servo Manipulation: Let Robots Swing a Hammer
Yingtian Xu, Ziya Wang
Robotic IntelligenceTime Series
🎯 What it does: Proposed a closed-loop tactile servo control strategy for controlling hammer slippage in a rigid two-finger gripper to achieve high-dynamic mechanical nail gun-like hammering.
High-fidelity Model and Nonlinear Model Predictive Control for Flip Maneuvers of Tailless Flapping-Wing Robots
Qingcheng Guo, Josie Hughes
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Developed a high-fidelity model and designed a nonlinear model predictive control (NMPC) algorithm to achieve 360-degree rolling of a tailless flapping-wing robot.
High-Precision and High-Efficiency Trajectory Tracking for Excavators Based on Closed-Loop Dynamics
Ziqing Zou, Yue Wang
OptimizationRobotic IntelligenceWorld Model
🎯 What it does: Proposed and verified a trajectory tracking method called EfficientTrack, combining model-based learning with closed-loop dynamics to address the nonlinear dynamics of excavators and improve tracking accuracy and efficiency.
High-Precision Parallel Manipulation of Multi-Particle System Using Optoelectronic Tweezers
Shunxiao Huang, Lin Feng
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: A photonic-electric micro-manipulator system with multi-particle parallel control was developed, integrating computer vision technology to achieve precise parallel manipulation of tens of particles.
High-Precision Pose Estimation of Medical Targets Using a Distortion Compensation Model for Robotic Surgical Navigation *
Weifeng Kong, Yiming Wang
Pose EstimationRobotic IntelligenceBiomedical Data
🎯 What it does: A method for medical target pose estimation using pre-calibration and Perspective-n-Point (PnP) algorithm, combined with a distortion compensation model and auto-exposure polarized vision to improve localization accuracy.
High-Precision Tracking of Time-Varying Trajectories for Microsurgical Robots in Constrained Environments
Yu-Peng Zhai, Yawen Deng
Robotic IntelligenceBiomedical Data
🎯 What it does: A hybrid control framework combining fuzzy adaptive sliding mode control and radial basis function (RBF) neural networks was developed and verified for high-precision time-varying trajectory tracking of minimally invasive surgical robots in constrained environments.
High-Precision Transformer-Based Visual Servoing for Humanoid Robots in Aligning Tiny Objects
J. Xue, Shiwu Zhang
Robotic IntelligenceTransformerImage
🎯 What it does: Propose a vision-based framework that utilizes the Transformer to accurately estimate and control the relative position between the handheld tool and the target object, achieving precise alignment of micro-objects.
High-Stiffness Path Planning for 7-DOF Cable-Driven Manipulators in Single and Dual-Arm Configurations
Shunxiang Pang, Weiwei Shang
OptimizationRobotic Intelligence
🎯 What it does: Proposed a motion planning framework to enhance the stiffness of single-arm and dual-arm cable-driven humanoid robotic arms; the single-arm approach combines dynamic obstacle avoidance with posture optimization to maximize end-effector stiffness; the dual-arm approach develops a coupled stiffness model to address dynamics between arms, enhancing performance in coordinated tasks.
HiTail: Hierarchical Neural Planner for Adaptive and Flexible Long-Tail Trajectory Planning
Shenghong Zhang, Xiao Li
Autonomous DrivingExplainability and InterpretabilityImage
🎯 What it does: Proposes a hierarchical neural trajectory planner that utilizes bird's-eye view (BEV) grid input, generating spatial proposals by sampling interpretable reward maps in a two-stage process, followed by assigning learnable time-speed curves (Clothoid curves) to these proposals.
Homotopy-aware Multi-agent Navigation via Distributed Model Predictive Control
Haoze Dong, Zhongkui Li
Autonomous DrivingOptimization
🎯 What it does: Proposes a distributed trajectory planning framework that adopts homotopy class-based optimal path planning at the global level and model predictive control (MPC)-based trajectory optimization at the local level, with online replanning added to adapt to environmental changes.
HPLaw: Heterogeneous Parallel LiDARs for Adverse Weather in V2V
Yuhang Liu, Fei-Yue Wang
Autonomous DrivingKnowledge DistillationRepresentation LearningPoint CloudBenchmark
🎯 What it does: Propose the HPLaw framework, creating two new benchmarks, OPV2V-W and V2V4Real-W, to study sensor heterogeneity in V2V under adverse weather conditions, and enhance model robustness through self-knowledge distillation.
Human Demonstrations are Generalizable Knowledge for Robots
Guangyan Chen, Yufeng Yue
Knowledge DistillationRobotic IntelligenceLarge Language ModelVideoRetrieval-Augmented Generation
🎯 What it does: Proposes the DigKnow method, which converts frames from human demonstration videos into observational knowledge, further extracts action knowledge, and distills it into pattern knowledge containing task and object instances, forming hierarchical, generalizable knowledge. Subsequently, relevant knowledge is retrieved in different task or object instance scenarios, utilizing LLMs to plan and execute actions, and verifying and correcting the planning and execution results, thereby significantly improving task success rates.
Human Implicit Preference-Based Policy Fine-tuning for Multi-Agent Reinforcement Learning in USV Swarm
Hyeonjun Kim, Jinkyoo Park
Reinforcement Learning from Human FeedbackLarge Language Model
🎯 What it does: Proposes a reinforcement learning with human feedback (RLHF) method for unmanned surface vehicles (USV) swarms in multi-agent reinforcement learning (MARL), addressing the credit assignment problem and validating its effectiveness through a large language model (LLM) evaluator.
Human-guided robotic-assistance handheld continuum medical robot system
Fei Wang, Haojian Lu
Robotic Intelligence
🎯 What it does: Proposed a human-guided handheld continuum medical robot system (HRHC), which simulates intuitive manual operations and achieves robotic precision, enhancing surgeons' capabilities in minimally invasive surgery while maintaining portability.
Human-in-the-loop Learning for Adaptive Robot Manipulation using Large Language Models and Behavior Trees
Haotian Zhou, Huasong Min
Robotic IntelligenceReinforcement Learning from Human FeedbackTransformerLarge Language Model
🎯 What it does: Proposes a human-robot collaborative learning mechanism that leverages large language models to generate action knowledge through in-context learning, and refines and improves it via user feedback, thereby achieving adaptive robot operations based on behavior trees;
Human-In-the-loop Optimisation in Robot-Assisted Gait Training
Andreas Christou, S. Vijayakumar
OptimizationHyperparameter SearchRobotic IntelligenceBiomedical Data
🎯 What it does: Utilizing human-in-the-loop optimization (HILO) in robotic-assisted gait training, the study employs CMA-ES to continuously optimize the assist-as-needed controller of the lower limb exoskeleton, conducting two-day experiments and collecting data from six healthy participants.
Human-Inspired Planning and Control of Shotcrete Robots based on Dynamical Systems Mapping
Rui Wu, Aude Billard
Robotic Intelligence
🎯 What it does: Proposes a new planning strategy for jet concrete robots, covering both the spraying and surface polishing stages, capable of handling flat or complex curved surface targets, and achieving adaptive planning in the presence of disturbances.
Human-Inspired Soft Anthropomorphic Hand System for Neuromorphic Object and Pose Recognition Using Multimodal Signals
Fengyi Wang, Gordon Cheng
ClassificationRecognitionPose EstimationSpiking Neural NetworkMultimodality
🎯 What it does: Developed a multimodal sensorized soft humanoid hand, converting its perceptual signals into pulse sequences via bio-inspired encoding, then using spiking neural networks (SNNs) for object and pose recognition, while introducing a differential neuron model to enhance material classification;
Human-Robot Co-Transportation using Disturbance-Aware MPC with Pose Optimization
Al Jaber Mahmud, Xuan Wang
OptimizationRobotic Intelligence
🎯 What it does: Proposing a dual-robot collaborative transportation control algorithm based on MPC and posture optimization
Human-Robot Collaborative SLAM-XR
Mohamad Karim Yassine, Daniel C. Asmar
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Proposes a collaborative centralized 3D mapping and localization framework that integrates SLAM and XR, capable of fusing local maps from multiple heterogeneous robots and supporting human intervention at multiple levels;
Human-Robot Cooperative Heavy Payload Manipulation based on Whole-Body Model Predictive Control
Ning Wang, Tianwei Zhang
OptimizationRobotic Intelligence
🎯 What it does: Developed a framework integrating a collaborative controller and a whole-body model predictive controller (MPC) for human-robot collaboration in heavy object transportation.
Human-Robot Shared Visual Servoing Based on Game Theory
Zitai Fang, Lijun Han
OptimizationRobotic Intelligence
🎯 What it does: Proposed a human-robot shared visual servoing framework based on game theory
Humanoid Whole-Body Locomotion on Narrow Terrain via Dynamic Balance and Reinforcement Learning
Weiji Xie, Xuelong Li
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed a full-body walking algorithm based on dynamic balance and reinforcement learning, utilizing ZMP-driven rewards and task-driven rewards, enabling bipedal robots to walk on extremely narrow terrain and unexpected obstacles using only proprioception
Hybrid Control Approach for Walking-Assembly Integrated Space Robots in On-Orbit Assembly*
Darran A.J. Douglas, Xinle Yan
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed a hybrid control strategy combining MPC and RL for a 7-degree-of-freedom walking assembly integrated space robot
Hybrid Data-Driven Predictive Control for Robust and Reactive Exoskeleton Locomotion Synthesis
Kejun Li, A. D. Ames
Robotic Intelligence
🎯 What it does: Proposed and verified the hybrid data-driven predictive control framework (HDDPC), achieving robust and adaptive gait for exoskeletons by simultaneously planning foot contact schedules and continuous domain trajectories.
Hybrid Data-Model-Driven External Force Estimation for Manipulators via Generalized Momentum-Based Third-Order Observer*
Haohao Zhang, Zhongyi Ren
Robotic Intelligence
🎯 What it does: Proposes a hierarchical fusion framework, first using a multi-layer perceptron neural network (MLPNN) to compensate for joint force residuals, and then estimating the external forces acting on the robotic arm through a third-order external force observer based on generalized momentum.