IROS 2025 Papers — Page 20
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1984 papers
UVS: A Novel Underwater Vehicle with Integrated VCMS-Thrusters Hybrid Architecture for Enhanced Attitude Regulation
Suohang Zhang, Yanhu Chen
Optimization
🎯 What it does: Proposed and implemented a hybrid attitude regulation architecture combining a variable center of mass system (VCMS) with thrusters—UVS;
VaLID: Verification as Late Integration of Detections for LiDAR-Camera Fusion
Vanshika Vats, James Davis
Object DetectionAutonomous DrivingImageMultimodalityPoint Cloud
🎯 What it does: Propose a model-adaptive late fusion method called VaLID to verify the reliability of each LiDAR detection box.
VAPO: Visibility-Aware Keypoint Localization for Efficient 6DoF Object Pose Estimation
Ruyi Lian, Haibin Ling
Pose EstimationComputational EfficiencyImage
🎯 What it does: Proposed a 3D keypoint localization method based on visibility-driven importance weights, and constructed the VAPO framework to enhance the accuracy of 6DoF pose estimation.
VCADNet: Vision-based Circular Accessible Depth Prediction for UGV Perception
Tao Zhang, Wei Zhang
Depth EstimationAutonomous DrivingConvolutional Neural NetworkContrastive LearningImagePoint Cloud
🎯 What it does: Proposed a vision-based circular reachable depth prediction network, VCADNet, for unmanned ground vehicle (UGV) perception.
VDTF-ACT: ACT-based Multimodal Space Fine Manipulation Method with Visual Depth Tactile Fusion
S. Lang, Zhiqiang Ma
Robotic IntelligenceTransformerMultimodality
🎯 What it does: Studied a multimodal perception-based space micro-manipulation method to enhance the precise operation performance of satellite robots in low-gravity environments.
VecNav: Vector Goal Robot Navigation from In-the-wild Videos
Ruixiang Cao, Jun Morimoto
Robotic IntelligenceTransformerDiffusion modelVideo
🎯 What it does: Proposed the VecNav method, which trains a monocular navigation model through self-supervised learning using uncalibrated, human-captured in-the-wild videos;
Vehicle Drifting Planning and Control Framework for Flexible U-turns in Space-limited Environments
Shuaicong Yang, Mengyin Fu
Autonomous DrivingOptimization
🎯 What it does: Propose a planning and control framework that utilizes drift maneuvers to achieve flexible U-turns in space-constrained environments, incorporating a dual-track 3-DoF vehicle model, nonlinear optimization planner, and multi-layer tracking controller.
VERAGMIL: Virtual Environment for Scooping Granular Foods with Imitation Learning Models
Amanuel Ergogo, P. Korzeniowski
Robotic IntelligenceRecurrent Neural NetworkVideo
🎯 What it does: Constructed the VERAGMIL framework, integrating a high-fidelity simulator and VR interface, to record human demonstrations and train a robot-assisted feeding system for handling tasks involving granular foods such as rice and beans; evaluated imitation learning models including BC, BC-RNN, and BCQ;
VerifyLLM: LLM-Based Pre-Execution Task Plan Verification for Robots
Danil S. Grigorev, Aleksandr I. Panov
Robotic IntelligenceLarge Language ModelText
🎯 What it does: Using a large language model (LLM) for pre-execution verification of robot task planning, first converting natural language instructions into linear temporal logic (LTL), then using LLM reasoning to evaluate the logical consistency of action sequences and identify potential gaps.
Versatile Demonstration Interface: Toward More Flexible Robot Demonstration Collection
Michael Hagenow, Julie Shah
Robotic IntelligenceImage
🎯 What it does: Proposes a multifunctional demonstration interface (VDI) installable on collaborative robots, which captures three common demonstration methods through vision, force sensing, and state tracking, and validates its usability via user studies conducted at the Manufacturing Innovation Center.
VertiSelector: Automatic Curriculum Learning for Wheeled Mobility on Vertically Challenging Terrain
Tong Xu, Xuesu Xiao
Robotic IntelligenceReinforcement LearningPhysics Related
🎯 What it does: Proposed VertiSelector (VS), an adaptive terrain sampling-based auto-curriculum learning framework to enhance the efficiency and generalization capability of robot reinforcement learning on vertical challenging terrains, validated in the VW-Chrono1 simulation environment and Verti-4-Wheeler physical platform.
VET: A Visual-Electronic Tactile System for Immersive Human-Machine Interaction
Cong Zhang, Wenbo Ding
Robotic Intelligence
🎯 What it does: Designed and implemented a visual-electronic tactile (VET) system that integrates visual-based tactile sensors (VBTS) with electrical stimulation membranes through a screen printing process, enabling bidirectional tactile interaction. Experimental validation was conducted in scenarios such as finger electrical stimulation sensitive areas, interactive games, and robotic arm teleoperation.
ViaTac: A High-Resolution Piezoresistive Tactile Sensor Array with Conformal Contact Surface for Shape Reconstruction
Yanjun Du, Dongyan Xu
Robotic Intelligence
🎯 What it does: Designed and implemented a low-cost, easy-to-manufacture piezoresistive tactile sensor array called ViaTac, which utilizes flexible printed circuit (FPC) via holes as electrodes and employs stretchable materials for surface encapsulation, achieving high spatial resolution (64 cm⁻²) and capable of forming good conformity with object surfaces for shape reconstruction tasks.
VibeCheck: Using Active Acoustic Tactile Sensing for Contact-Rich Manipulation
Kaidi Zhang, M. Ciocarlie
ClassificationRecognitionPose EstimationRobotic IntelligenceAudio
🎯 What it does: Built and verified a dual piezoelectric fingertip gripper based on active acoustic tactile sensing, used to acquire acoustic responses during object contact-rich operations, achieving object classification, grasp position estimation, internal structure pose estimation, and external contact type identification.
Vibration-Assisted Hysteresis Mitigation for Achieving High Compensation Efficiency
Myeongbo Park, Minho Hwang
OptimizationComputational EfficiencyRobotic IntelligenceConvolutional Neural Network
🎯 What it does: Proposed applying controlled vibrations in the tendon sheath mechanism (TSM) to reduce friction and dead zones, thereby decreasing hysteresis and improving trajectory tracking accuracy.
Vibration-Aware Trajectory Optimization for Mobile Robots in Wild Environments via Physics-Informed Neural Network
Ao Xu, Yu Hu
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Proposed a physics-informed trajectory planning method that considers the vibration effects of complex nonlinear suspension systems;
Vibration-Based Energy Metric for Restoring Needle Alignment in Autonomous Robotic Ultrasound
Zhongyu Chen, K. W. S. Au
Robotic IntelligenceBiomedical DataUltrasound
🎯 What it does: Propose an energy measurement method based on mechanical vibration for automatic correction when the ultrasound imaging plane and needle insertion plane are mismatched, and develop corresponding control strategies to reposition the probe in translational and rotational directions.
Vibration-induced Friction Modulation to Enable Controlled Sliding for In-hand Manipulation
Shambhuraj Mane, B. Çalli
Robotic Intelligence
🎯 What it does: By embedding vibration modules on a mechanical finger, the study investigated surface vibration to regulate friction for achieving controllable sliding, designed an activation state selection method, and demonstrated controllable sliding and rotation of held objects on a two-fingered robot hand.
Vibrotactile Sensing for Detecting Misalignments in Precision Manufacturing
Kevin Zhang, Oliver Kroemer
Robotic Intelligence
🎯 What it does: Proposes a new method for robotic manipulation in high-mix low-volume manufacturing environments using tactile vibration sensors
Video-Rate 4D OCT Segmentation Based on Motion-Aware Probabilistic A-Scan Sampling
Shervin Dehghani, Nassir Navab
SegmentationConvolutional Neural NetworkBiomedical Data
🎯 What it does: Proposed an efficient segmentation network based on motion-aware probabilistic A-scan sparse sampling and using only 1D convolutions, achieving real-time 4D OCT volume semantic segmentation;
View-aware Decomposition and Unification for Fast Ground-to-Aerial Person Search
Qifei Wang, Yongsheng Gao
RetrievalContrastive LearningImage
🎯 What it does: Proposes a perspective-aware decomposition and unification (VADU) framework to model perspective differences in ground-to-air person search.
ViewActive: Active viewpoint optimization from a single image
Jiayi Wu, Y. Aloimonos
OptimizationRobotic IntelligenceConvolutional Neural NetworkImage
🎯 What it does: Propose ViewActive, a lightweight active viewpoint optimization method based on a single image, which guides robots to acquire the optimal observation angle through the 3D Viewpoint Quality Field (VQF).
VIMS: A Visual-Inertial-Magnetic-Sonar SLAM System in Underwater Environments
Bingbing Zhang, Wen Xu
Simultaneous Localization and MappingMultimodality
🎯 What it does: Proposed and implemented an underwater SLAM system called VIMS, which integrates visual, inertial, magnetometer, and low-cost single-beam sonar, aiming to address the challenges of scale estimation and loop detection in traditional visual-inertial systems under water.
Vine4Spine: A Steerable Tip-Growing Robot with Contact Force Estimation for Navigation in the Spinal Subarachnoid Space
Zicong Wu, S. Sadati
Robotic IntelligenceBiomedical Data
🎯 What it does: Designed and evaluated a miniature steerable flipping robot for spinal subarachnoid navigation, capable of real-time contact force estimation and reducing interaction with anatomical structures.
VINS-MLD2: Monocular Visual-Inertial SLAM With Multi-level Detector and Descriptor
X. Nian, Yong Chen
Convolutional Neural NetworkSimultaneous Localization and MappingOptical FlowImageMultimodality
🎯 What it does: Proposes a visual-inertial SLAM system called VINS-MLD2 based on multi-level detectors and descriptors, designs an efficient deep feature extraction network, and introduces matching fusion and adaptive matching strategies.
VISC: mmWave Radar Scene Flow Estimation using Pervasive Visual-Inertial Supervision
Kezhong Liu, Shengkai Zhang
Optical FlowMultimodality
🎯 What it does: Propose a millimeter-wave radar scene flow estimation framework supervised by extensive visual-inertial (VI) sensor data, including a drift-eliminated rigid transformation estimator and an optical-millimeter wave supervision extraction module.
Vision Guided Cable Installation in Constraint Environments Utilizing Parametric Curve Representation
Xin Jiang, Wei Ran
Robotic IntelligenceImage
🎯 What it does: A vision-based cable installation method is proposed, combining end-effector trajectory regulation based on potential field control with shape deformation servo, and determining feasible obstacle-avoidance shape curves through a planner;
Vision-Based Contact Wrench Estimation in Human-Robot Interaction
Mohammad Farajtabar, Marie Charbonneau
Pose EstimationDepth EstimationRobotic IntelligenceRecurrent Neural NetworkImagePoint Cloud
🎯 What it does: A vision-based multi-contact physical human-robot interaction torque estimation method is proposed, which uses RGB-D sensors to detect 3D hand positions, identify contact points, and employs a generalized momentum observer to distinguish joint torque from external torque; an LSTM network is used to compensate for uncertainties in unmodeled dynamics.
Vision-Based Cooperative MAV-Capturing-MAV
Canlun Zheng, Shiyu Zhao
Autonomous DrivingOptimizationRobotic IntelligenceImage
🎯 What it does: A vision-based multi-drone collaborative capture system was designed and implemented, utilizing distributed state estimation and control. Multiple pursuer drones optimize trajectories via MPC and execute them using an SO(3) low-level controller. After meeting capture conditions, the system automatically deploys a flying net to intercept the target drone.
Vision-Based Force Feedback System Using Moiré Patterns
Jinrun Zuo, Takeshi Takaki
Robotic IntelligenceImage
🎯 What it does: Propose a force feedback system based on visual processing and Moiré patterns, which utilizes image processing for real-time measurement and provides force feedback
Vision-Based Tactile Sensor Using Light-Conductive Plate for Enhanced Force Sensing Capability
Zhitong Liu, Xin Jiang
Computational EfficiencyRobotic IntelligenceImage
🎯 What it does: Developed a vision-based tactile sensor (VBTS) that uses a light guide plate and silicone membrane to monitor deformation at the contact point, achieving high-precision estimation of contact position and force (micrometer-level resolution) through image processing and deep learning.
Vision-Driven 2D Supervised Fine-Tuning Framework for Bird’s Eye View Perception
Lei He, Keqiang Li
SegmentationDepth EstimationAutonomous DrivingSupervised Fine-TuningImage
🎯 What it does: We propose a fine-tuning method for BEV perception networks based on visual 2D semantic perception to enhance the model's generalization ability on new scene data.
Vision-Guided Loco-Manipulation with a Snake Robot
Adarsh Salagame, Alireza Ramezani
Object DetectionPose EstimationDepth EstimationRobotic IntelligenceConvolutional Neural NetworkImage
🎯 What it does: Developed and integrated a vision-based walking-manipulation pipeline for the snake robot COBRA at Northeastern University, achieving real-time object detection, 6-DOF pose estimation, and closed-loop transportation;
Vision-Language Guided Adaptive Robot Action Planning: Responding to Intermediate Results and Implicit Human Intentions
Weihao Cai, Nobutaka Shimada
Robotic IntelligencePrompt EngineeringVision Language ModelVision-Language-Action ModelMultimodality
🎯 What it does: Generate feasible robot action sequences using visual prompts and Visual Language Models (VLM) to support human-robot collaboration while simultaneously estimating human intent.
Vision-Language Navigation with Continual Learning for Unseen Environments
Zhiyuan Li, Hong Qiao
Autonomous DrivingVision-Language-Action Model
🎯 What it does: Propose the Vision-Language Navigation with Continuous Learning (VLNCL) framework, and enable the VLN agent to learn in new environments while retaining old knowledge through Dual-SR (Dual-Cycle Scene Replay).
VisLanding: Monocular 3D Perception for UAV Safe Landing via Depth-Normal Synergy
Zhuoyue Tan, Liaoni Wu
SegmentationDepth EstimationConvolutional Neural NetworkSupervised Fine-TuningImage
🎯 What it does: Developed a VisLanding framework based on monocular depth-normal joint prediction for safe drone landing.
VISO-Grasp: Vision-Language Informed Spatial Object-centric 6-DoF Active View Planning and Grasping in Clutter and Invisibility
Yitian Shi, R. Rayyes
Pose EstimationRepresentation LearningRobotic IntelligenceTransformerVision Language ModelMultimodality
🎯 What it does: Designed a vision-language system called VISO-Grasp, which utilizes foundation models for spatial reasoning and active view planning, constructs instance-centric spatial relationship representations, and achieves 6-DoF grasping in heavily occluded environments through a multi-view uncertainty-driven grasping fusion mechanism;
Visual Anomaly Detection for Reliable Robotic Implantation of Flexible Microelectrode Array
Yitong Chen, Shan Yu
Anomaly DetectionRobotic IntelligenceTransformerImage
🎯 What it does: Developed an image anomaly detection framework based on a microscope camera for monitoring four checkpoints during the robotic implantation of flexible microelectrode arrays (microneedles, microelectrode probes, connection results, implantation sites).
Visual Localization with Offline Google Satellite Map-Assisted for Ground Vehicles in GNSS-Denied Environment
Jibo Wang, Zheng Fang
Pose EstimationAutonomous DrivingSimultaneous Localization and MappingImage
🎯 What it does: Propose a visual localization framework that utilizes offline Google satellite maps to address vehicle localization challenges in environments where GNSS signals are weak or unavailable; introduce a learning-based ground-satellite map feature matching method, propose a cross-perspective pose selection approach to construct two pose uncertainty models, and integrate them with classical SLAM methods;
Visual Loop Closure Detection Through Deep Graph Consensus
Martin Buchner, Abhinav Valada
Autonomous DrivingGraph Neural NetworkSimultaneous Localization and MappingImage
🎯 What it does: Proposes LoopGNN, a graph neural network architecture that utilizes visual similarity keyframe cliques for loop closure detection;
Visual-Haptic Model Mediated Teleoperation for Remote Ultrasound
David G. Black, Marco Esposito
Robotic IntelligenceImageMultimodalityUltrasound
🎯 What it does: Propose a robotic teleoperated ultrasound system that real-time re-slices and renders pre-acquired ultrasound images with the assistance of a visual-tactile model to alleviate the impact of time delay on operations, and tested on 15 volunteers.
ViT-VS: On the Applicability of Pretrained Vision Transformer Features for Generalizable Visual Servoing
Alessandro Scherl, Jos'e Grac'ia-Rodr'iguez
Robotic IntelligenceTransformerImage
🎯 What it does: Propose a visual servoing method that utilizes a pre-trained vision Transformer for semantic feature extraction.
VLIN-RL: A Unified Vision-Language Interpreter and Reinforcement Learning Motion Planner Framework for Robot Dynamic Tasks
Zewu Jiang, Chenyi Si
Robotic IntelligenceReinforcement LearningVision Language ModelMultimodality
🎯 What it does: Proposed the VLIN-RL framework, integrating the Vision-Language Interpreter (VLIN) with a reinforcement learning-based motion planner to enable real-time adjustment of subtasks based on visual feedback during task execution, thereby enhancing the real-time performance and robustness of robotic dynamic tasks.
VLM Can Be a Good Assistant: Enhancing Embodied Visual Tracking with Self-Improving Vision-Language Models
Kui Wu, Fangwei Zhong
Object TrackingVision Language Model
🎯 What it does: Propose a self-improving framework that integrates visual language models (VLM) with entity visual tracking (EVT) for active visual tracking and to address tracking failure recovery issues.
VLM See, Robot Do: Human Demo Video to Robot Action Plan via Vision Language Model
Beichen Wang, Chen Feng
Robotic IntelligenceVision Language ModelVision-Language-Action ModelVideoBenchmark
🎯 What it does: Propose a system called SeeDo that can generate natural language task plans for robot execution from long-term human demonstration videos.
VLM-Empowered Multi-Mode System for Efficient and Safe Planetary Navigation
Sinuo Cheng, Liang Ding
Autonomous DrivingVision Language Model
🎯 What it does: Proposed a multi-mode navigation system based on vision-language models to achieve efficient and safe autonomous navigation for planetary rovers.
VMTS: Vision-Assisted Teacher-Student Reinforcement Learning for Multi-Terrain Locomotion in Bipedal Robots
Fu Chen, Bo Zhou
Robotic IntelligenceReinforcement LearningMixture of ExpertsImage
🎯 What it does: Propose a hybrid expert teacher-student network reinforcement learning strategy, combining terrain selection strategy with teacher strategy, and introduce an alignment loss between teacher and student networks, verifying its feasibility and robustness on the Limx Dynamic P1 bipedal robot across various terrains.
VoxEKF-RIO: A 4D Radar Inertial Odometry Based on Incremental Voxel Map and Iterated Kalman Filter
Jiawei Shen, Lei Sun
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposed a 4D radar inertial odometry method called VoxEKF-RIO based on incremental voxel grids and iterative Kalman filtering.
VPOcc: Exploiting Vanishing Point for 3D Semantic Occupancy Prediction
Junsu Kim, Kyungdon Joo
Autonomous DrivingImageBenchmark
🎯 What it does: Proposes the VPOcc framework, which addresses the scale inconsistency between 2D and 3D at both pixel and feature levels by leveraging vanishing points (VP), enabling camera-based 3D semantic occupancy prediction.
VRobotix: A Scalable and Cost-Effective Virtual-Reality-Based Robotic Manipulation Dataset Generation Framework
Xinmin Fang, Zhengxiong Li
GenerationData SynthesisRobotic IntelligenceMultimodality
🎯 What it does: Developed a VR-based human-robot interaction framework called VRobotix, which generates scalable, low-cost robotic grasping and manipulation datasets in physically accurate simulation environments, and provides a replay module and imitation learning module;
VSG-SLAM:A Dense Visual Semantic SLAM with Gaussian Splatting
Wenyuan Tong, Limin Zeng
Pose EstimationOptimizationGaussian SplattingSimultaneous Localization and Mapping
🎯 What it does: Proposes a dense visual semantic SLAM framework that combines high-dimensional semantic features with explicit 3D Gaussian scattering scene representation
VSLAM-LAB: A Comprehensive Framework for Visual SLAM Methods and Datasets
Alejandro Fontan, Michael Milford
Simultaneous Localization and MappingImageBenchmark
🎯 What it does: Proposed the VSLAM-LAB unified framework to simplify VSLAM algorithm development, evaluation, and deployment, achieving algorithm compilation, configuration, automatic dataset download and preprocessing, standardized experimental design, and assessment through a single command-line interface.
VTAO-BiManip: Masked Visual-Tactile-Action Pre-training with Object Understanding for Bimanual Dexterous Manipulation
Zheng Sun, Jiming Chen
Robotic IntelligenceReinforcement LearningVision-Language-Action ModelMultimodality
🎯 What it does: Proposed the VTAO-BiManip framework, which integrates vision-tactile-action pre-training with object understanding, utilizes bimanual motion data to predict future actions and object pose/size, and achieves bimanual coordination through two-phase curriculum reinforcement learning.
Wave-Aware Control of Workspace-Constrained Shipboard Robots for Motion Compensation in Rough Seas
Lingda Kong, Zhenyu Gao
Robotic Intelligence
🎯 What it does: Propose a hierarchical planning and model predictive control framework, enabling ship-mounted robots to execute highly dynamic motions for motion compensation in rough sea conditions.
WAVE: Worm Gear-based Adaptive Variable Elasticity for Decoupling Actuators from External Forces
Moses Gladson Selvamuthu, Kazutoshi Tanaka
Robotic Intelligence
🎯 What it does: Designed and experimentally verified the variable stiffness actuator WAVE based on spiral gears, achieving separation between the drive motor and external forces, and continuously adjusting joint stiffness through spring pre-compression length.
Weakly-supervised VLM-guided Partial Contrastive Learning for Visual Language Navigation
Ruoyu Wang, Lina Yao
RecognitionVision Language ModelContrastive LearningMultimodality
🎯 What it does: Propose Weakly-Supervised Partial Contrastive Learning (WPCL), which enhances object recognition in Vision-and-Language Navigation (VLN) by effectively integrating pre-trained Vision-Language Model (VLM) knowledge without requiring fine-tuning of the VLM.
Wearable Roller Rings to Augment In-Hand Manipulation through Active Surfaces
H. Webb, Kaiyu Hang
Robotic Intelligence
🎯 What it does: Designed and tested a wearable roller ring (Roller Ring), enabling in-hand manipulation of objects of any shape without lifting fingers by installing roller rings at different angles on the palm.
Wearing a Robotic Hand to Feel 3D Force Feedback: Analysis and Virtual Reality Application of the Hand-in-Hand System
Nicolas Kosanovic, J. Vaz
Pose EstimationRobotic Intelligence
🎯 What it does: Convert a 3D-printed robotic hand costing approximately $500 into a wearable exoskeleton system (Hand-in-Hand), achieving 3D force feedback at the fingertips and validating its performance in VR human-robot interaction experiments.
WebRTC and 5G Based Remote Control System for a Vascular Intervention Robot
Sheng Cao, Junbo Ge
Robotic IntelligenceImage
🎯 What it does: Developed a remote vascular interventional robotic control system based on WebRTC and 5G networks
Weight Regression for a Generalized Motion Primitive Formulation in Cooperative Hand Placement Tasks with Upper-Limb Prostheses
Hongjun Cai, Nitish V. Thakor
Robotic IntelligenceSequential
🎯 What it does: Propose a bio-inspired shared control strategy that utilizes dynamic movement primitives (DMP) to adaptively generate smooth trajectories from a stationary position to any reachable point, and associates DMP driving functions with Cartesian positions through weight regression.
WFDA: Wavelet-Based Frequency Decomposition and Aggregation for Underwater Object Detection
Xueting Liu, Shuxiang Guo
Object DetectionImage
🎯 What it does: Propose a Wavelet-Based Frequency Decomposition and Aggregation Network (WFDA) for underwater target detection, utilizing wavelet transforms for feature frequency decomposition and aggregation.
WHALES: A Multi-Agent Scheduling Dataset for Enhanced Cooperation in Autonomous Driving
Siwei Chen, Sheng Zhou
Autonomous DrivingOptimizationAgentic AIPoint CloudBenchmark
🎯 What it does: Proposed a large-scale V2X dataset called WHALES and designed a scheduling algorithm named CAHS based on historical perspective coverage
What Really Matters for Robust Multi-Sensor HD Map Construction?
Xiaoshuai Hao, Rong Yin
Autonomous DrivingMultimodalityPoint Cloud
🎯 What it does: Proposed data augmentation, a multi-modal fusion module, and a modal dropout training strategy to enhance the robustness and accuracy of multi-sensor HD map construction.
Where to Wait: Postponing the Decision About Waiting Locations in Multi-Agent Path Finding
David Zahrádka, Miroslav Kulich
OptimizationBenchmark
🎯 What it does: Proposed the Partial Safe Interval (PSI) mechanism, allowing delayed decision-making on waiting positions in multi-agent path planning and extending the Prioritized Safe Interval Path Planning (PSIPP) algorithm;
Whole-Body Admittance Control of Anti-Saturation for Quadruped Manipulators with Impact Force Observer
Fenghao Lin, Yunjiang Lou
OptimizationRobotic Intelligence
🎯 What it does: Designed and implemented a full-body impedance control framework that estimates collision forces without force/torque sensors using a novel external impact force observer, and limits joint torque through set-valued feedback control to prevent saturation, enhancing safety and compliance.
Whole-Body Impedance Control of a Humanoid Robot Based on Human-Human Demonstration for Human-Robot Collaboration
Chenzui Li, Fei Chen
Robotic IntelligenceBiomedical Data
🎯 What it does: Propose a whole-body impedance control method applied to the collaborative dual-arm robot CURI, which can adapt robot behavior according to human motion and follow trajectories learned from human-human demonstration;
Whole-Body Stabilization of Wheeled Bipedal Robots via Decoupled Control of Wheels and Legs
Jechan Jeon, Yonghwan Oh
Robotic Intelligence
🎯 What it does: Proposed a full-body control framework based on a decoupled architecture, utilizing a TWIP template to control only wheel motion, allowing the full-body controller to focus exclusively on leg dynamics.
Wireless Collaborative Inference Acceleration Based on Distillation for Weed Detection and Instance Segmentation
Rongjiao Li, Lei Mu
Object DetectionSegmentationComputational EfficiencyKnowledge DistillationConvolutional Neural NetworkImageAgriculture Related
🎯 What it does: Proposes a wireless collaborative inference framework for deep learning instance segmentation in resource-constrained weed robots, improving feature fusion, RPN (Soft-NMS), Mask branch, employing knowledge distillation for model compression, and providing two-stage optimal partitioning points and resource-aware dynamic optimization algorithms, finally verifying its feasibility in a drone-server system.
WiTAH A*: Winding-Constrained Anytime Heuristic Search for a Pair of Tethered Robots
Xingjian Xue, Sze Zheng Yong
OptimizationRobotic Intelligence
🎯 What it does: Proposes a variant of the Anytime Hybrid A* algorithm that first generates a fast but suboptimal path, then gradually optimizes it to solve the shortest path problem with curvature constraints for two tethered robots under minimum curling angle constraints.
WLuav: An Air-Ground Robot with High Ground Adaptability and Trajectory Tracking Performance
Shijie Huang, Yaonan Wang
Robotic IntelligenceBenchmark
🎯 What it does: Proposed a ground-air robot named WLuav based on a five-bar wheel-legged structure, designed a hierarchical adaptive agile controller and a mode switching strategy based on a support force solver, and subsequently validated its performance through experiments and benchmark comparisons.
WMNav: Integrating Vision-Language Models into World Models for Object Goal Navigation
Dujun Nie, Long Chen
Robotic IntelligenceVision Language ModelWorld ModelMultimodality
🎯 What it does: Proposed WMNav, a navigation framework based on a world model, which utilizes a vision-language model to predict decision outcomes and constructs memory feedback to the policy module;
World Models for Anomaly Detection during Model-Based Reinforcement Learning Inference
Fabian Domberg, Georg Schildbach
Anomaly DetectionReinforcement LearningWorld Model
🎯 What it does: Leverage the discrepancy between model predictions and actual observations to continuously monitor during inference, detecting anomalies in the state space and triggering safety measures such as emergency shutdowns in unfamiliar regions.
Would you let a humanoid play storytelling with your child? A usability study on LLM-powered narrative Humanoid-Robot Interaction
Maria Lombardi, Agnieszka Wykowska
Robotic IntelligenceTransformerLarge Language Model
🎯 What it does: Proposes an iCub robot attention enhancement framework that integrates social cue recognition, generative models (such as ChatGPT), and appropriate social behaviors, validated through interactive narrative tasks.
Wrench-Guided and Velocity-Field-Based Geometric Impedance Control*
Yuancan Huang, Da Hong
Robotic Intelligence
🎯 What it does: A geometric impedance control method based on velocity field was designed, generating impedance-related velocity fields on SE(3) to characterize local interaction behavior.
YO-CSA-T: A Real-time Badminton Tracking System Utilizing YOLO Based on Contextual and Spatial Attention
Yuan Lai, Chengxi Zhu
Object DetectionObject TrackingConvolutional Neural NetworkVideo
🎯 What it does: Developed a real-time badminton racket trajectory tracking system that integrates the YO-CSA detection network, stereo vision 3D mapping, and a trajectory prediction and compensation module to achieve end-to-end 2D-to-3D tracking;
YOLO-MARL: You Only LLM Once for Multi-Agent Reinforcement Learning
Zhuang Yuan, Fei Miao
Large Language ModelReinforcement Learning
🎯 What it does: Propose the YOLO-MARL framework, which utilizes a large language model (LLM) to generate strategies, state explanations, and planning functions in one interaction, followed by training decentralized multi-agent policies.
ZBOT: A Novel Modular Robot Capable of Active Transformation from Snake to Bipedal Configuration through RL
Nanlin Zhou, Yanhe Zhu
Robotic IntelligenceReinforcement Learning
🎯 What it does: Designed a modular robot ZBOT capable of actively transforming from a snake-like morphology to a bipedal morphology, and trained and verified its stance gait in the IsaacSim/Lab simulation environment using reinforcement learning.
Zero Shot Domain Adaptive Semantic Segmentation by Synthetic Data Generation and Progressive Adaptation
Jun Luo, Yang Liu
SegmentationData SynthesisDomain AdaptationDiffusion modelImageText
🎯 What it does: This paper proposes a zero-shot domain adaptation semantic segmentation method called SDGPA, which generates synthetic training images in the target domain style using text descriptions and performs progressive adaptation.
Zero-Shot Peg Insertion: Identifying Mating Holes and Estimating SE(2) Poses with Vision-Language Models
M. Yajima, Rei Kawakami
RecognitionPose EstimationVision Language ModelMultimodality
🎯 What it does: Propose a zero-shot plug insertion framework that utilizes a vision-language model (VLM) to identify matching holes and estimate their SE(2) pose, completing the insertion task under the assumption of a known plug pose and planar surface.
Zero-Shot Semantic Segmentation for Robots in Agriculture
Y. Linn, C. Stachniss
SegmentationAnomaly DetectionRobotic IntelligenceAgriculture Related
🎯 What it does: Propose a zero-shot anomaly segmentation method for unsupervised semantic segmentation of crops and weeds; extract features using a pre-trained base model, construct a bag feature representation for crop features, and model the feature space of crop plants as a hypersphere; during inference, plants falling within this hypersphere are classified as crops, while the rest are considered weeds.
Zero-Shot Temporal Interaction Localization for Egocentric Videos
Erhang Zhang, Hesheng Wang
Vision Language ModelVideo
🎯 What it does: Proposed a zero-shot temporal interaction localization method called EgoLoc.
Zoned Artificial Repulsion: Path Planning Through Local Minima for Multiple-Robot Dexterous Micromanipulation
Tala Dannawi Aissaoui, Redwan Dahmouche
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented the Zoned Artificial Repulsion (ZAR∗) algorithm for path planning in multi-robot micro-manipulation tasks, combining an improved artificial potential field (APF) with A*, using graph search to achieve globally optimal paths after partitioning the configuration space into regions.
ZS-Puffin: Design, Modeling and Implementation of an Unmanned Aerial-Aquatic Vehicle with Amphibious Wings
Zhenjiang Wang, Wubin Wang
Robotic Intelligence
🎯 What it does: Proposed a two-ambient unmanned aerial-aquatic vehicle (UAAV) equipped with amphibious wings, which are modified from a fixed-wing structure, allowing single-degree-of-freedom pitch motion without additional components, capable of generating lift in the air and achieving flapping propulsion underwater, and equipped with an artificial central pattern generator (CPG) to enhance the smoothness of flapping motion. The paper presents the prototype, design details, and implementation.
λ: A Benchmark for Data-Efficiency in Long-Horizon Indoor Mobile Manipulation Robotics
Ahmed Jaafar, Stefanie Tellex
Robotic IntelligenceVision-Language-Action ModelMultimodalityBenchmark
🎯 What it does: Proposed the LAMBDA (λ) benchmark to evaluate models' data efficiency in language-conditioned, multi-room, multi-floor, long-horizon, and pick-and-place tasks, and compared end-to-end learning methods with modular neuro-symbolic approaches.