ICRA 2024 Papers — Page 3
IEEE International Conference on Robotics and Automation · 1760 papers
AnyOKP: One-Shot and Instance-Aware Object Keypoint Extraction with Pretrained ViT
Fangbo Qin, Shan Yu
Pose EstimationRepresentation LearningTransformerImage
🎯 What it does: Proposed a single-shot instance-aware object keypoint extraction method called AnyOKP based on pre-trained Vision Transformer, which can extract keypoints from multi-instance images of any category after learning a support image.
APACE: Agile and Perception-aware Trajectory Generation for Quadrotor Flights
Xinyi Chen, Shaojie Shen
OptimizationRobotic IntelligenceImage
🎯 What it does: Proposes the APACE framework, which designs a perception-aware trajectory generation method considering feature matching for aggressive quadrotor flights.
Approximate Multiagent Reinforcement Learning for On-Demand Urban Mobility Problem on a Large Map
Daniel Garces, Stephanie Gil
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: Propose an approximate multi-agent rollout two-phase algorithm for automated multi-vehicle soldier path planning in large-scale urban environments, achieving stability and near-optimal strategies;
Approximating Robot Configuration Spaces with few Convex Sets using Clique Covers of Visibility Graphs
P. Werner, Russ Tedrake
Robotic Intelligence
🎯 What it does: Proposes an efficient method to approximately cover complex robot configuration spaces using cluster coverage based on visibility graphs
Are you a robot? Detecting Autonomous Vehicles from Behavior Analysis
Fabio Maresca, Xavier Pérez Costa
ClassificationData SynthesisAutonomous DrivingVideoTabular
🎯 What it does: Designed and implemented a framework based on camera images and vehicle state information for automatically detecting the presence of autonomous vehicles on roads.
ARIS 1.0: An Autonomous Multitasking Medical Service Robot for Hospital Environments
D.M.A.P. Dunuwila, X. Tan
Robotic IntelligenceSimultaneous Localization and MappingBiomedical Data
🎯 What it does: Developed ARIS 1.0, a multi-task medical service robot supporting remote healthcare and medical education, featuring a three-wheel omnidirectional mobile platform, movable neck, and facial expression capabilities.
ArrayBot: Reinforcement Learning for Generalizable Distributed Manipulation through Touch
Zhengrong Xue, Huazhe Xu
Robotic IntelligenceReinforcement Learning
🎯 What it does: Designed the ArrayBot system with a 16×16 vertical sliding column array, and utilized RL algorithms to automatically learn control policies through tactile observations, achieving support, perception, and distributed manipulation of desktop objects.
Articulated Object Manipulation with Coarse-to-fine Affordance for Mitigating the Effect of Point Cloud Noise
Suhan Ling, Hao Dong
Robotic IntelligencePoint Cloud
🎯 What it does: Proposed a coarse-to-fine operability learning pipeline to alleviate the impact of real-world point cloud noise on articulated object manipulation.
ASAP: Automated Sequence Planning for Complex Robotic Assembly with Physical Feasibility
Yunsheng Tian (Massachusetts Institute of Technology), Wojciech Matusik (Massachusetts Institute of Technology)
Robotic IntelligenceGraph Neural Network
🎯 What it does: Propose ASAP (Automated Sequence Planning), a physics-based planning method for automatically generating physically feasible assembly sequences applicable to complex products with various shapes.
ASGrasp: Generalizable Transparent Object Reconstruction and 6-DoF Grasp Detection from RGB-D Active Stereo Camera
Jun Shi, He Wang
Object DetectionData SynthesisPose EstimationConvolutional Neural NetworkImagePoint Cloud
🎯 What it does: Proposes ASGrasp, a 6-DoF grasp detection network using RGB-D active stereo cameras, which employs a dual-layer learning-based stereo network to achieve transparent object reconstruction, enabling material-agnostic grasping.
ASP-LED: Learning Ambiguity-Aware Structural Priors for Joint Low-Light Enhancement and Deblurring
Jing Ye, Zhiyong Zhang
RestorationTransformerImage
🎯 What it does: Proposes a novel fuzzy perception network ASP-LED, achieving joint learning of low-light enhancement and deblurring using structural prior (high-frequency and edges).
ASPIRe: An Informative Trajectory Planner with Mutual Information Approximation for Target Search and Tracking
Kangjie Zhou, Chang Liu
Object Tracking
🎯 What it does: Propose an adaptive particle filter tree trajectory planning method based on mutual information reward approximation for mobile target search and tracking in limited visibility.
Assessing Reputation to Improve Team Performance in Heterogeneous Multi-Robot Coverage
Mela C. Coffey, Alyssa Pierson
Robotic Intelligence
🎯 What it does: Studied using robot reputation to evaluate historical performance in heterogeneous multi-robot coverage tasks for dynamically adjusting work area sizes, thereby improving overall team performance.
Assessment and Benchmarking of XoNLI: a Natural Language Processing Interface for Industrial Exoskeletons
Olmo A. Moreno-Franco, J. Ortiz
Robotic IntelligenceBenchmarkAudio
🎯 What it does: Evaluation and benchmarking of the voice interaction interface XoLab Natural Language Interface for industrial active exoskeletons.
Assisting Group Discussions Using Desktop Robot Haru
Fei Tang, Guangliang Li
Robotic IntelligenceReinforcement Learning
🎯 What it does: Developed a desktop robot system named Haru to assist group discussions, comprising three modules: dialogue assistance, dialogue balancing, and autonomous gaze.
Asymptotically-Optimal Multi-Robot Visibility-Based Pursuit-Evasion
Nicholas M. Stiffler, Jason M. O'Kane
OptimizationRobotic IntelligenceGraph
🎯 What it does: Propose a new asymptotically optimal multi-robot visibility pursuit-evasion algorithm that utilizes three hierarchical graph data structures to generate cooperative motion strategies.
Asynchronous Distributed Smoothing and Mapping via On-Manifold Consensus ADMM
Daniel McGann, Michael Kaess
OptimizationSimultaneous Localization and Mapping
🎯 What it does: Developed a fully distributed, asynchronous, and generic Consensus SLAM backend optimization algorithm called MESA.
Asynchronous Task Plan Refinement for Multi-Robot Task and Motion Planning
Yoonchang Sung, Peter Stone
OptimizationRobotic Intelligence
🎯 What it does: Model the plan refinement problem in multi-robot task and motion planning as a hybrid constraint satisfaction problem, and propose an algorithm based on sequential heuristics and implicit time and road network representations.
ATPPNet: Attention based Temporal Point cloud Prediction Network
Kaustab Pal, K. M. Krishna
GenerationAutonomous DrivingRecurrent Neural NetworkTransformerPoint CloudTime Series
🎯 What it does: Proposed and implemented ATPPNet, an attention mechanism-based temporal point cloud prediction network for predicting future point cloud sequences acquired by LiDAR sensors;
Attention-Based Cloth Manipulation from Model-free Topological Representation
Kevin Galassi, J. Renders
Robotic IntelligenceTransformerPoint CloudMesh
🎯 What it does: Designed and implemented an attention-based neural network for smoothing fabrics with a single-arm robot in both simulated and real environments.
Attitude Control for Morphing Quadrotor through Model Predictive Control with Constraints*
Na Zhao, Yantao Shen
Optimization
🎯 What it does: Proposed and applied constrained model predictive control (MPC) to address the attitude control problem of deformable quadrotors.
Augmenting Lane Perception and Topology Understanding with Standard Definition Navigation Maps
Katie Z Luo, Marco Pavone
Autonomous DrivingTransformerGraph
🎯 What it does: Investigated the effectiveness of standard definition maps in real-time lane topology understanding, and proposed a framework to integrate standard definition maps into online map prediction.
Augmenting Tactile Simulators with Real-like and Zero-Shot Capabilities
Osher Azulay, A. Sintov
Image TranslationData SynthesisGenerative Adversarial NetworkImage
🎯 What it does: Propose a bidirectional generative adversarial network called SightGAN to learn the mapping between real and simulated tactile images, and achieve high-resolution realistic synthesis of tactile images by precisely reconstructing small contact traces through additional background and contact pattern loss.
AutoExplorers: Autoencoder-Based Strategies for High-Entropy Exploration in Unknown Environments for Mobile Robots
Lennart Puck, R. Dillmann
Robotic IntelligenceAuto EncoderImage
🎯 What it does: This paper proposes using an autoencoder to embed environmental data from drones or satellites, selecting the next high-entropy exploration target based on distance metrics in the embedding space to improve detection efficiency in unknown environments.
AutoFusion: Autonomous Visual Geolocation and Online Dense Reconstruction for UAV Cluster
Yizhu Zhang, Lin Chen
Pose EstimationOptimizationGraph Neural NetworkSimultaneous Localization and MappingImage
🎯 What it does: Proposed a real-time multi-drone dense reconstruction system named AutoFusion, which can achieve autonomous visual localization and map fusion under conditions of lost GNSS signals and weak visibility
Automated Assembly by Two-Fingered Microhand for Fabrication of Soft Magnetic Microrobots
Yue Zhao, T. Arai
Robotic Intelligence
🎯 What it does: Developed an automated assembly system based on a two-finger micro-hand for preparing magnetic soft micromachines.
Automated Non-invasive Analysis of Motile Sperms Using Cross-scale Guidance Network
Wei Dai, Jun Liu
SegmentationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: Proposed and implemented a cross-scale guidance network (CSG) for non-invasive precise segmentation of tiny sperm, and conducted morphological and kinematic parameter analysis based on segmentation results.
Automated Sperm Immobilization with a Clinically-Compatible and Compact XYZ Stage
Haocong Song, Yu Sun
Object TrackingRobotic IntelligenceBiomedical Data
🎯 What it does: Developed a three-dimensional positioning platform capable of automatically locating and fixing human sperm, and achieved sperm tracking and positioning through a visual servo controller.
Automated Sperm Morphology Analysis Based on Instance-Aware Part Segmentation
Wenyuan Chen, Yu Sun
SegmentationConvolutional Neural NetworkBiomedical Data
🎯 What it does: Proposes an automatic sperm morphology analysis method based on instance-aware partial segmentation, and provides an automatic measurement scheme for tail morphology.
Automated Surgical Knot Tying on Mini-Incision with Micro-Suture based on Dual-Arm Nanorobot under Stereo Microscope
Yujie Jiang, Song Liu
Robotic Intelligence
🎯 What it does: Proposed and implemented a system for automatic knotting of micro sutures through incisions using dual-arm nanorobots under a stereomicroscope, with motion trajectory planning, experimental evaluation, and mechanical strength testing conducted.
Automated Testing of Spatially-Dependent Environmental Hypotheses through Active Transfer Learning
Nicholas Harrison, Salah Sukkarieh
Domain AdaptationData-Centric Learning
🎯 What it does: Developed an active transfer learning method capable of real-time evaluation of spatially related environmental hypotheses, utilizing existing prior data to improve sampling efficiency
Automatic Captioning based on Visible and Infrared Images
Yan Wang, Huaping Liu
GenerationVision Language ModelImageMultimodality
🎯 What it does: This paper addresses the complementarity between visible light images and infrared images, proposing an automatic caption generation model for RGBIR image fusion and developing a wearable environmental assistance system; meanwhile, a dataset of 3510 pairs of RGB-IR images was collected and annotated for model training and evaluation.
Automatic Configuration of Multi-Agent Model Predictive Controllers based on Semantic Graph World Models
K. Vos, R. V. D. Molengraft
Autonomous DrivingOptimizationRobotic IntelligenceGraph Neural NetworkWorld ModelGraph
🎯 What it does: Propose a shared semantic map architecture for dynamically building and configuring multi-robot model predictive controllers (MPC), addressing multi-robot navigation problems in shared environments.
Automatic Loading of Unknown Material with a Wheel Loader Using Reinforcement Learning
D. Eriksson, Marcus Geimer
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: Studied the use of reinforcement learning to fine-tune an automated controller for a full-scale 24-ton wheeled loader to handle unknown material loading, enabling adaptation from trained rock materials to unknown materials such as gravel;
Automatic Trust Estimation From Movement Data in Industrial Human-Robot Collaboration Based on Deep Learning
Matthias Rehm, Kasper Hald
Robotic IntelligenceTime Series
🎯 What it does: Investigate whether the trust level of users towards robots can be estimated in real-time through observed motion data in industrial human-robot collaboration, and collect data from two collaborative tasks for experimentation
Automatically designing robot swarms in environments populated by other robots: an experiment in robot shepherding
David Garzón-Ramos, Mauro Birattari
Robotic IntelligenceReinforcement Learning
🎯 What it does: Automatically generate control software for robot shepherding tasks through automatic modular design and neuroevolutionary techniques, enabling a group of automatically designed shepherding robots to interact and coordinate with pre-programmed sheep robot groups.
Autonomous 3D Exploration in Large-Scale Environments with Dynamic Obstacles
Emil Wiman, Fredrik Heintz
Autonomous DrivingOptimizationRobotic IntelligenceSimultaneous Localization and MappingBenchmark
🎯 What it does: Propose a Dynamic Autonomous Exploration Planner (DAEP) that actively considers dynamic obstacles in planning and leverages the advantages of dynamic environments to enhance exploration efficiency;
Autonomous aerial perching and unperching using omnidirectional tiltrotor and switching controller
Dongjae Lee, H. J. Kim
Robotic Intelligence
🎯 What it does: Developed a lightweight fully actuated tilt-rotor aircraft capable of autonomous hovering, landing, and takeoff on vertical magnetic surfaces, and designed a single servo motor + magnet landing/takeoff module, combined with a switching controller to achieve smooth transition from free flight to landing/takeoff.
Autonomous and Teleoperation Control of a Drawing Robot Avatar
Lingyun Chen, Sami Haddadin
OptimizationRobotic IntelligenceImage
🎯 What it does: Proposes an automated control framework that enhances the visual feedback quality of eye-hand cameras by computing approximately optimal end-effector poses, thereby reducing user workload in remote drawing tasks, and validates its effectiveness through user studies.
Autonomous Apple Fruitlet Sizing with Next Best View Planning
Harry Freeman, G. Kantor
Robotic IntelligenceGraph Neural NetworkAgriculture Related
🎯 What it does: Proposed a next best view planning method for autonomous measurement of apple granules
Autonomous Exploration of Unknown 3D Environments Using a Frontier-Based Collector Strategy
Ivan D. Changoluisa Caiza, Tamara Petrović
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Utilizing UAVs for autonomous exploration, a global exploration method combining the frontier collector strategy is proposed, enabling the creation of a complete map.
Autonomous Field-of-View Adjustment Using Adaptive Kinematic Constrained Control with Robot-Held Microscopic Camera Feedback
Hung-Ching Lin, Kanako Harada
Robotic IntelligenceConvolutional Neural NetworkImage
🎯 What it does: Developed an autonomous control method to restrict the robot's handheld microscope camera movement within a specified field of view.
Autonomous Implicit Indoor Scene Reconstruction with Frontier Exploration
Jing Zeng, Jiming Chen
OptimizationNeural Radiance FieldSimultaneous Localization and MappingImage
🎯 What it does: Proposes an adaptive viewpoint planning method that combines frontier exploration tasks with implicit surface uncertainty reconstruction based on color uncertainty
Autonomous Mapless Navigation on Uneven Terrains
Hassan Jardali, Lantao Liu
OptimizationRobotic IntelligencePoint Cloud
🎯 What it does: A local perception model based on sparse Gaussian processes is proposed, which learns terrain elevation and extracts feasible sub-goals using point cloud observations without relying on maps. Subsequently, a cost function prioritizing robot roll and pitch angle constraints is used to select safe sub-goals in real-time, guiding the robot to navigate to the target position.
Autonomous Overhead Powerline Recharging for Uninterrupted Drone Operations
Viet Duong Hoang, E. Ebeid
Robotic IntelligencePhysics Related
🎯 What it does: Proposed a fully autonomous drone system capable of self-charging and long-duration continuous operation near power transmission lines.
Autonomous Perching on Flat Surfaces for Free-Flying Robots with Gecko Adhesive Gripper
D. Hirano, Tony G. Chen
Robotic Intelligence
🎯 What it does: A passive mechanism and control method using a gecko-inspired adhesive gripper for free-flying robots is proposed, achieving automatic docking on flat surfaces in microgravity environments.
Autonomous Quilt Spreading for Caregiving Robots
Yuchun Guo, Xin Jiang
SegmentationPose EstimationDepth EstimationRobotic IntelligenceConvolutional Neural NetworkTransformerImage
🎯 What it does: This paper proposes an automatic baby blanket laying strategy for care robots, which can promptly and accurately re-cover the blanket after the baby shifts it during sleep.
Autonomous robotic re-alignment for face-to-face underwater human-robot interaction*
Demetrious T. Kutzke, Junaed Sattar
Pose EstimationDepth EstimationRobotic IntelligenceImage
🎯 What it does: A stereo vision system for underwater human-robot interaction was developed, utilizing stereo image pairs for 3D reconstruction and machine learning for human joint localization. Subsequently, a coordinate system was constructed to encode the human's facing direction relative to the camera, enabling automated setpoint calculation, which serves as input for image-based visual servo control.
Autonomous UAV Mission Cycling: A Mobile Hub Approach for Precise Landings and Continuous Operations in Challenging Environments
Alexander Moortgat-Pick, Sami Haddadin
Robotic Intelligence
🎯 What it does: Propose the concept of a mobile center, installing a small landing platform on a robotic arm to enable drones to achieve precise landing and continuous task cycles without modification; simultaneously, maintenance (e.g., battery replacement) and direct relaunch are performed via the mobile center.
AutoTAMP: Autoregressive Task and Motion Planning with LLMs as Translators and Checkers
Yongchao Chen, Chuchu Fan
Robotic IntelligenceTransformerLarge Language ModelPrompt EngineeringText
🎯 What it does: Propose the AutoTAMP method, which leverages large language models (LLMs) to translate natural language task descriptions into intermediate task representations using few examples, and then combines traditional task and motion planning (TAMP) algorithms to jointly generate task and motion plans; meanwhile, autoregressive re-prompting techniques are employed to automatically detect and correct syntactic and semantic errors to improve task completion rates.
AYDIV: Adaptable Yielding 3D Object Detection via Integrated Contextual Vision Transformer
Tanmoy Dam, Mir Feroskhan
Object DetectionAutonomous DrivingTransformerImagePoint Cloud
🎯 What it does: Proposes the AYDIV framework, which integrates a three-stage alignment process to enhance long-range object detection performance under LiDAR and camera fusion.
Barrier Functions Inspired Reward Shaping for Reinforcement Learning
Nilaksh Nilaksh, Shishir N Y Kolathaya
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed and implemented a safety-oriented reward shaping framework based on barrier functions, evaluated in simulation environments such as CartPole, Ant, Humanoid, and on the Unitree Go1 quadruped robot.
BaSAL: Size-Balanced Warm Start Active Learning for LiDAR Semantic Segmentation
Jiarong Wei, Holger Caesar
SegmentationAutonomous DrivingPoint Cloud
🎯 What it does: Proposed and implemented BaSAL, a size-balanced warm-start active learning model for LiDAR semantic segmentation.
Bayesian Constraint Inference from User Demonstrations Based on Margin-Respecting Preference Models
D. Papadimitriou, Daniel S. Brown
Robotic IntelligenceReinforcement Learning from Human Feedback
🎯 What it does: Propose a constraint reasoning framework based on Bayesian methods to infer robot environmental constraints from user demonstrations according to a preference model.
Bayesian-Guided Evolutionary Strategy with RRT for Multi-Robot Exploration
Shuge Wu, Zhongliang Zhao
OptimizationRobotic Intelligence
🎯 What it does: Proposes an adaptive RRT tree growth strategy for frontier point detection, as well as a Bayesian-guided evolutionary strategy (BGE) for efficient task allocation.
BeBOP - Combining Reactive Planning and Bayesian Optimization to Solve Robotic Manipulation Tasks
J. Styrud, Christian Smith
OptimizationRobotic Intelligence
🎯 What it does: Proposes a method called BeBOP that combines reactive planning with Bayesian optimization for generating behavior trees, used to rapidly configure robot manipulation tasks;
BEE-Net: Bridging Semantic and Instance with Gated Encoding and Edge Constraint for Efficient Panoptic Segmentation
Xinyang Huang, Jiamao Li
SegmentationComputational EfficiencyConvolutional Neural NetworkImage
🎯 What it does: Designed a new panoptic, instance, and semantic bridging network, and proposed the Gated Encoding module and edge-aware consistency constraints to improve panoptic segmentation;
Behavior Tree Capabilities for Dynamic Multi-Robot Task Allocation with Heterogeneous Robot Teams
G. Heppner, R. Dillmann
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented a dynamic task allocation system based on behavior trees, where users can specify tasks using BT, and the system dynamically assigns tasks to heterogeneous robot teams during runtime.
Behavioral-based circular formation control for robot swarms
Jesus Bautista, Héctor García de Marina
OptimizationRobotic Intelligence
🎯 What it does: Designed a behavior-based circular formation control algorithm to coordinate robot swarms around curved trajectory paths, avoiding collisions and achieving high-density formations.
Belief Scene Graphs: Expanding Partial Scenes with Objects through Computation of Expectation
M. A. Saucedo, G. Nikolakopoulos
Robotic IntelligenceGraph Neural NetworkGraph
🎯 What it does: Propose Belief Scene Graphs (BSG), extending partial 3D scene graphs by computing expected values to support robot task planning.
Benchmarking Classical and Learning-Based Multibeam Point Cloud Registration
Li Ling, A. Wåhlin
Pose EstimationPoint CloudBenchmark
🎯 What it does: Benchmarking classical ICP family methods and learning-based registration methods on a semi-synthetic multibeam sonar point cloud dataset.
Benchmarking Multi-Robot Coordination in Realistic, Unstructured Human-Shared Environments
Lukas Heuer, Martin Magnusson
Robotic IntelligenceBenchmark
🎯 What it does: A benchmark framework for multi-robot coordination algorithms in real-world, unstructured, and human-shared environments is proposed, with experimental evaluations conducted in two environments on three centralized coordination methods (two MAPF algorithms and one loosely coupled method based on precedence constraints).
Better Monocular 3D Detectors with LiDAR from the Past
Yurong You, Kilian Q. Weinberger
Object DetectionAutonomous DrivingPoint Cloud
🎯 What it does: Proposed the AsyncDepth framework, which provides asynchronous LiDAR features to monocular 3D detectors during inference using unlabeled historical LiDAR scans, thereby enhancing detection performance.
Bevel-Tip Needle Deflection Modeling, Simulation, and Validation in Multi-Layer Tissues
Yanzhou Wang, I. Iordachita
Biomedical Data
🎯 What it does: Proposed and validated a mechanics-based 2D bevel-tip needle model, and simulated multi-control inputs along the needle length and complete three degrees of freedom (3-DOF) planar motion through real-time finite element method (FEM) simulation.
BEVUDA: Multi-geometric Space Alignments for Domain Adaptive BEV 3D Object Detection
Jiaming Liu, Shanghang Zhang
Object DetectionDomain AdaptationAutonomous DrivingPoint Cloud
🎯 What it does: Proposes a multi-space aligned teacher-student framework (MATS), achieving domain adaptation for cross-domain bird's-eye view (BEV) 3D object detection through depth-aware teacher (DAT) and geometry space aligned student (GAS).
Bi-KVIL: Keypoints-based Visual Imitation Learning of Bimanual Manipulation Tasks
Jianfeng Gao, Tamim Asfour
Robotic IntelligenceVideo
🎯 What it does: Propose the Bi-KVIL method, extending K-VIL to bimanual manipulation, extracting Hybrid Master-Slave Relationships (HMSR), bimanual coordination strategies, and sub-symbolic task representations, enabling learning of fine-grained bimanual manipulation tasks from a small number of human demonstration videos.
Bi2Lane: Bi-Directional Temporal Refinement with Bi-Level Feature Aggregation for 3D Lane Detection
Chengxin Li, Jun Li
Autonomous DrivingConvolutional Neural NetworkVideo
🎯 What it does: Built an end-to-end monocular 3D lane detection framework called Bi2Lane, leveraging bidirectional temporal information from consecutive frames and dual-layer feature aggregation to achieve more robust lane detection.
Bicode: A Hybrid Blinking Marker System for Event Cameras
Takuya Kitade, Michita Imai
Pose EstimationImage
🎯 What it does: Proposed and experimentally tested a hybrid blinking marker system called Bicode, which integrates 2D markers with blinking LEDs into a single unit specifically designed for event cameras.
Bigraph Matching Weighted with Learnt Incentive Function for Multi-Robot Task Allocation
Steve Paul, Souma Chowdhury
OptimizationRobotic IntelligenceGraph Neural NetworkTransformerReinforcement LearningGraph
🎯 What it does: Propose a heuristic learning framework based on Graph Reinforcement Learning for large graph matching in multi-robot task allocation.
Bio-Inspired Pupal-Mode Actuator with Ultra-Crossing Capability for Soft Robots
Zhenxing Wang, Hao Liu
Robotic IntelligenceBiomedical Data
🎯 What it does: Designed and tested a pupal-mode soft actuator based on pupal-stage movement, achieving peristaltic-like motion through internal air chambers, and verified its performance in modal crossover environments using a gastroscope robot based on this actuator.
Bio-inspired visual relative localization for large swarms of UAVs
Martin Křížek, M. Saska
Robotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Proposed a vision-based UAV swarm relative localization method
Bionic Soft Fingers with Hybrid Variable Stiffness Mechanisms for Multimode Grasping
Xiangbo Wang, Bin Fang
Robotic Intelligence
🎯 What it does: Designed and implemented a bio-inspired soft finger (BSF) with a hybrid variable stiffness mechanism, incorporating an integrated shape memory alloy skeleton, particle jamming technology, and a water-cooled circulation system, enabling rapid operation and multi-mode grasping; simultaneously integrated with a liquid metal strain sensor (METT) to achieve real-time monitoring of fingertip bending angles.
Block-Map-Based Localization in Large-Scale Environment
Yixiao Feng, Guyue Zhou
Autonomous DrivingOptimizationPoint Cloud
🎯 What it does: Propose a localization system based on Block Maps, which reduces computational load for large-scale maps through Block Map generation and switching strategies, and provides global localization using Branch-and-Bound Search and dynamic sliding window graph optimization.
Body Velocity Estimation in a Leg–Wheel Transformable Robot without A Priori Knowledge of Leg–Wheel Ground Contacts
Pei-Chun Huang, Pei-Chun Lin
Robotic IntelligenceTime SeriesSequential
🎯 What it does: Proposes a method for simultaneously estimating velocity and ground contact state for deformable leg-wheel robots using a combination of inertial measurement unit (IMU) and encoder data, without relying on prior ground contact information.
Boosting Offline Reinforcement Learning for Autonomous Driving with Hierarchical Latent Skills
Zenan Li, Hang Zhao
Autonomous DrivingReinforcement LearningAuto Encoder
🎯 What it does: Proposes a skill-based framework that utilizes Variational Autoencoder (VAE) to learn driving skills from offline demonstrations, and employs these skills as actions to enhance the performance of offline reinforcement learning in long-term planning.
Boundary Factors for Seamless State Estimation between Autonomous Underwater Docking Phases
Aldo Terán Espinoza, Jakob Kuttenkeuler
OptimizationRobotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: For autonomous underwater docking targeting dynamic objects, a factor graph optimization-based SLAM framework is proposed, along with the design of boundary factors to achieve seamless transitions in state estimation.
Brain-Inspired Hyperdimensional Computing in the Wild: Lightweight Symbolic Learning for Sensorimotor Controls of Wheeled Robots
Hyukjun Kwon, Yeseong Kim
Robotic IntelligenceSupervised Fine-TuningReinforcement Learning
🎯 What it does: Proposed a lightweight framework called ReactHD based on hypervector computing for robot sensor-action perception-driven learning and control
Breaking Data Silos: Cross-Domain Learning for Multi-Agent Perception from Independent Private Sources
Jinlong Li, Hongkai Yu
Domain AdaptationAutonomous DrivingPoint Cloud
🎯 What it does: Proposes the Feature Distribution-aware Aggregation (FDA) framework to address the distribution gap issue caused by independent private data in multi-agent perception systems, aiming to break data silos.
Bridging the Sim-to-Real Gap with Dynamic Compliance Tuning for Industrial Insertion
Xiang Zhang, Hui Li
Domain AdaptationRobotic Intelligence
🎯 What it does: Proposed a framework composed of Force Planner and Gain Tuner, achieving robust manipulation of industrial insertion tasks through dynamic adjustment of compliance control gain using only simulated data.
Bridging Zero-shot Object Navigation and Foundation Models through Pixel-Guided Navigation Skill
Wenzhe Cai, Hao Dong
Autonomous DrivingRobotic IntelligenceTransformerLarge Language ModelVision-Language-Action ModelImage
🎯 What it does: This paper proposes the Pixel-guided Navigation technique, which achieves zero-shot goal navigation using pixel targets, constructs a pure RGB navigation strategy, and realizes long-distance goal planning through an LLM planner.
BroadBEV: Collaborative LiDAR-camera Fusion for Broad-sighted Bird’s Eye View Map Construction
Minsu Kim, Sunwook Choi
Autonomous DrivingTransformerImagePoint Cloud
🎯 What it does: Propose a BEV fusion framework named BroadBEV, aiming to enhance the accuracy of camera BEV estimation and address the full BEV space defects caused by the sparsity of LiDAR point clouds;
Building User Proficiency in Piloting Small Unmanned Aerial Vehicles (sUAV)
Siya Kunde, Brittany A. Duncan
🎯 What it does: This paper proposes a training method based on two experiments for proficiency assessment and training of small unmanned aerial vehicle (UAV) operators
C2FDrone: Coarse-to-Fine Drone-to-Drone Detection using Vision Transformer Networks
S. V. Rebbapragada, Vineeth N. Balasubramanian
Object DetectionTransformerImage
🎯 What it does: Proposed a coarse-to-fine UAV-UAV detection strategy based on vision transformers, achieving real-time detection.
CalibFormer: A Transformer-based Automatic LiDAR-Camera Calibration Network
Yuxuan Xiao, Yanyong Zhang
Pose EstimationAutonomous DrivingTransformerImagePoint Cloud
🎯 What it does: Proposed an end-to-end automatic LiDAR-camera calibration network called CalibFormer, achieving high-resolution representations and accurate calibration parameter estimation by aggregating multi-layer camera and LiDAR image features, using multi-head correlation modules, and employing a Transformer architecture.
CalliRewrite: Recovering Handwriting Behaviors from Calligraphy Images without Supervision
Yuxuan Luo, Zhouhui Lian
Robotic IntelligenceReinforcement LearningImage
🎯 What it does: Propose a coarse-to-fine process to enable the robotic arm to recover feasible writing sequences and achieve precise control under unsupervised annotation conditions for different calligraphy images.
Camera Relocalization in Shadow-free Neural Radiance Fields
Shiyao Xu, Guyue Zhou
Pose EstimationOptimizationNeural Radiance FieldImage
🎯 What it does: Propose a two-stage pipeline to perform illumination and shadow normalization on images for camera relocalization, and implement scene representation based on hash-encoded NeRF to improve pose optimization efficiency; meanwhile, address gradient noise in grid-based NeRF by introducing a redesigned truncated dynamic low-pass filter (TDLF) and numerical gradient averaging techniques for smoothing.
CAMInterHand: Cooperative Attention for Multi-View Interactive Hand Pose and Mesh Reconstruction
Guwen Han, Jiming Chen
Pose EstimationConvolutional Neural NetworkImageMesh
🎯 What it does: Proposes a multi-view collaborative attention method called CAMInterHand for interactive hand pose and mesh reconstruction.
Campus Map: A Large-Scale Dataset to Support Multi-View VO, SLAM and BEV Estimation
J. Ross, Richard Bowden
Autonomous DrivingSimultaneous Localization and MappingImagePoint CloudBenchmark
🎯 What it does: Propose the Campus Map large-scale multi-camera dataset, containing 2M images, GPS, and 64-beam LiDAR scans, and provide 40k semantic Bird's Eye View (BEV) maps and a simulation environment.
Capacitive Origami Sensing Modules for Measuring Force in a Neurosurgical, Soft Robotic Retractor
Daniel Van Lewen, S. Russo
Robotic Intelligence
🎯 What it does: Developed a soft capacitor-based origami sensing module (OSM) capable of measuring force during neurosurgical stretching and integrated it into a soft robotic manipulator.
CAPE: Corrective Actions from Precondition Errors using Large Language Models
S. S. Raman, Stefanie Tellex
Robotic IntelligenceTransformerLarge Language ModelTextChain-of-Thought
🎯 What it does: Propose a method called CAPE for action generation based on large language models, which corrects preconditions errors and improves planning quality through few-shot reasoning, enabling embodied agents to better recover from failures when executing natural language commands.
CAPT: Category-level Articulation Estimation from a Single Point Cloud Using Transformer
Lian Fu, Takeshi Oishi
Pose EstimationTransformerPoint Cloud
🎯 What it does: Propose CAPT, which employs an end-to-end Transformer architecture to estimate joint parameters and states from a single point cloud, and introduces motion loss and a dual voting strategy.
CARTIER: Cartographic lAnguage Reasoning Targeted at Instruction Execution for Robots
Nikhil Kakodkar, Gregory Dudek
Robotic IntelligenceTransformerLarge Language ModelText
🎯 What it does: Using large language models (LLM) to parse natural conversational instructions in robot navigation through the CARTIER method.
CausalAgents: A Robustness Benchmark for Motion Forecasting
R. Roelofs, Wei Chai
Autonomous DrivingPoint CloudBenchmark
🎯 What it does: Constructed a new benchmark to evaluate and enhance the robustness of motion prediction models under non-causal perturbations, annotated causal agents in the Waymo Open Motion Dataset (WOMD), and perturbed the data by removing non-causal agents.
CaveSeg: Deep Semantic Segmentation and Scene Parsing for Autonomous Underwater Cave Exploration
A. Abdullah, Ioannis M. Rekleitis
SegmentationTransformerImageBenchmark
🎯 What it does: Developed CaveSeg—the first visual learning pipeline for AUV navigation in underwater caves, combining semantic segmentation and scene parsing, and prepared a comprehensive dataset with pixel annotations for navigation markers, obstacles, divers, and open areas.
Cellular-enabled Collaborative Robots Planning and Operations for Search-and-Rescue Scenarios
Arnau Romero, Xavier Pérez Costa
OptimizationRobotic Intelligence
🎯 What it does: Proposes a search and rescue framework for collaborative robots in cellular networks, taking area, fleet size, energy configuration, exploration rate, and response time as inputs, and outputting the minimum number of robots and their paths that meet task objectives.
Censible: A Robust and Practical Global Localization Framework for Planetary Surface Missions
J. Nash, Vandi Verma
Pose EstimationAutonomous DrivingRobotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Developed a global positioning algorithm for the Perseverance rover, utilizing a modified census transform to match onboard images with orbital maps.
Certifying Bimanual RRT Motion Plans in a Second
Alexandre Amice, Russ Tedrake
OptimizationRobotic Intelligence
🎯 What it does: Propose an efficient method for non-collision proof of piecewise polynomial motion planning in an algebraic reparameterized configuration space
CFDNet: A Generalizable Foggy Stereo Matching Network with Contrastive Feature Distillation
Zihua Liu, Masatoshi Okutomi
Depth EstimationKnowledge DistillationContrastive LearningImage
🎯 What it does: Proposes a framework based on Contrastive Feature Distillation (CFD) for foggy stereo matching;
Chained Flexible Capsule Endoscope: Unraveling the Conundrum of Size Limitations and Functional Integration for Gastrointestinal Transitivity
Sishen Yuan, Hongliang Ren
🎯 What it does: Designed and verified a chain-like flexible capsule endoscope (FCE), achieving chain-like flexibility through traditional rotational joint design and flexible material notches, addressing capsule size limitations while enhancing therapeutic functions.
Chance-Aware Lane Change with High-Level Model Predictive Control Through Curriculum Reinforcement Learning
Yubin Wang, Jun Ma
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: Proposed an opportunity-aware lane change strategy, implemented using high-level model predictive control (MPC) and curriculum reinforcement learning (CRL);