IROS 2024 Papers — Page 12
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers
PICaSo: A Collaborative Robotics System for Inpainting on Physical Canvas using Marker and Eraser
Shady Nasrat (University Of California Berkeley), Seung-Joon Yi
RestorationGenerationRobotic IntelligenceSupervised Fine-TuningVision-Language-Action ModelImageText
🎯 What it does: Proposed the PICaSo system, which enables collaborative graffiti and restoration on a real canvas through a robotic arm and brush, capable of automatically generating artworks based on user natural language prompts and positional instructions.
PickScan: Object discovery and reconstruction from handheld interactions
Vincent van der Brugge, Krishna Murthy Jatavallabhula
Object DetectionSegmentationGenerationDepth EstimationOptical FlowImage
🎯 What it does: Propose PickScan, a class-agnostic object discovery and reconstruction method based on handheld interaction.
Pilot Study for a Robot-Assisted Timed Up and Go Assessment*
Matthew Story, A. D. Nuovo
Pose EstimationRobotic IntelligenceConvolutional Neural NetworkVideo
🎯 What it does: Using the Turtlebot 4 equipped with MoveNet and various machine learning algorithms to perform robot-assisted evaluation of the Timed Up and Go (TUG) test.
PINN-Ray: A Physics-Informed Neural Network to Model Soft Robotic Fin Ray Fingers
Xing Wang, David Howard
Robotic IntelligencePhysics Related
🎯 What it does: Proposed and validated the PINN-Ray model for complex deformation modeling of soft Fin Ray grippers, compared with FEM, and provided an automated framework for design, manufacturing, and visual tracking.
PINSAT: Parallelized Interleaving of Graph Search and Trajectory Optimization for Kinodynamic Motion Planning
Ramkumar Natarajan, M. Likhachev
OptimizationGraph
🎯 What it does: Proposes PINSAT, a dynamic motion planning method that parallelizes graph search with trajectory optimization.
Planning for Long-Term Monitoring Missions in Time-Varying Environments
Alex Stephens, Nick Hawes
OptimizationRobotic IntelligenceReinforcement LearningAudio
🎯 What it does: An online planning method based on Bayesian Monte Carlo Tree Search and spatiotemporal Gaussian Processes is proposed to address the long-term monitoring problem for repeated finite-horizon tasks in the same environment.
Pneumatic bladder links with wide range of motion joints for articulated inflatable robots
Katsu Uchiyama, Ryuma Niiyama
Robotic Intelligence
🎯 What it does: Proposed a scalable robot composed of multiple pneumatic bladder linkages connected by Hillberry rolling contact joints, demonstrating its applications in load transportation with three-degree-of-freedom arms, two-degree-of-freedom arms, one-degree-of-freedom arms, and bipedal walking.
PoCo: Point Context Cluster for RGBD Indoor Place Recognition
Jing Liang, Arnie Sen
RetrievalConvolutional Neural NetworkMultimodality
🎯 What it does: Proposed an end-to-end RGB-D indoor scene recognition algorithm called PoCo, which is used to find the most matching location in the reference database for query frames.
Polaris: Open-ended Interactive Robotic Manipulation via Syn2Real Visual Grounding and Large Language Models
Tianyu Wang, Yanwei Fu
Pose EstimationRobotic IntelligenceTransformerLarge Language ModelVision-Language-Action ModelImageMultimodality
🎯 What it does: This paper proposes the Polaris framework, which combines GPT-4 with a visual localization model to achieve open interactive robotic arm operations in desktop scenarios, and introduces the Syn2Real end-to-end pose estimation pipeline.
PolyFit: A Peg-in-hole Assembly Framework for Unseen Polygon Shapes via Sim-to-real Adaptation
Geonhyup Lee, Kyoobin Lee
Domain AdaptationRobotic IntelligenceSupervised Fine-TuningTabular
🎯 What it does: Proposed the PolyFit framework, which utilizes force/torque (F/T) data through supervised learning to achieve 5-degree-of-freedom (DoF) plug-and-socket assembly, real-time estimation of external pose, and correction of plug posture to rectify errors;
Portable robot for needle insertion assistance to femoral artery
Zhuoqi Cheng, Olof Huldt
Robotic IntelligenceBiomedical DataUltrasound
🎯 What it does: Developed a portable robotic device for autonomous localization of the femoral artery and precise placement of a needle guide, assisting non-expert physicians in femoral artery puncture.
Pos2VPR: Fast Position Consistency Validation with Positive Sample Mining for Hierarchical Place Recognition
Dehao Zou, Zhuo Wang
Pose EstimationRetrievalComputational EfficiencySimultaneous Localization and MappingImage
🎯 What it does: Proposes a fast verification algorithm called PCLP based on local patch location consistency, and constructs a unified visual place recognition framework that integrates global feature aggregation with the PCLP verification module.
Pose Graph Optimization over Planar Unit Dual Quaternions: Improved Accuracy with Provably Convergent Riemannian Optimization
William D. Warke, Matthew T. Hale
Pose EstimationOptimizationSimultaneous Localization and Mapping
🎯 What it does: Proposes a Riemannian optimization algorithm based on the planar unit dual quaternion (PUDQ) manifold, improving the estimation of relative pose uncertainty in pose graph optimization (PGO) using a realistic correlated noise model.
Position Control of a Low-Energy C-Core Reluctance Actuator in a Motion System
M. A. Saaideh, M. Janaideh
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Proposed a position control system for a motion stage based on low-energy C-type magnetic reluctance actuators, reducing the required current through variable air gaps, linearizing dynamic behavior with a feedforward controller, and achieving tracking through state feedback.
Potential Field-Based Online Path Planning for Robust Cable Routing
Andrea Monguzzi, Paolo Rocco
Robotic Intelligence
🎯 What it does: This paper studies the path planning task of a single-arm robot sliding along an unknown curve on an elastic, deformable linear object (such as a cable or hose) that is constrained at both ends, performing collision-avoidance contour following and inserting a specified segment into a known fixture in the middle.
PP-TIL: Personalized Planning for Autonomous Driving with Instance-based Transfer Imitation Learning
Fangze Lin, F. Yu
Domain AdaptationAutonomous DrivingSupervised Fine-TuningReinforcement Learning
🎯 What it does: Proposed an instance-based transfer imitation learning method for personalized autonomous driving planning in complex urban environments.
Practical Framework for Path Representation and Following Control in Mobile Industrial Robots
Youngil Koh, MidEum Choi
OptimizationComputational EfficiencyRobotic Intelligence
🎯 What it does: Developed a practical framework for path representation and following control in mobile industrial robots, supporting the generation of drivable paths based on user-specified waypoints while considering allowable deviations and maximum curvature, and introducing a 'dock' attribute for each waypoint; simultaneously implemented a path-following control system that considers execution delay and computational efficiency, achieving robust tracking through speed planning and brake intervention using predictive information.
Pre-training on Synthetic Driving Data for Trajectory Prediction
Yiheng Li, Wei Zhan
Data SynthesisAutonomous DrivingRepresentation LearningAuto Encoder
🎯 What it does: Proposes a full-process solution based on HD map enhancement and trajectory synthesis, first generating synthetic driving data, then pre-training the trajectory prediction model using a masked autoencoder (MAE) to learn general representations.
PreAfford: Universal Affordance-Based Pre-Grasping for Diverse Objects and Environments
Kairui Ding, Hao Zhao
Representation LearningRobotic IntelligencePoint CloudMesh
🎯 What it does: Proposed a PreAfford pre-grasp planning framework based on point-level feasibility representation and relay training
Precise Pick-and-Place using Score-Based Diffusion Networks
Shihui Guo, Chun-Yi Lee
Robotic IntelligenceDiffusion modelScore-based ModelImage
🎯 What it does: Propose a coarse-to-fine cascade continuous pose diffusion method to enhance the accuracy of robotic grasping and placing.
Precise Well-plate Placing Utilizing Contact During Sliding with Tactile-based Pose Estimation for Laboratory Automation
S. Pai, Avinash Ummadisingu
Pose EstimationRobotic Intelligence
🎯 What it does: A robotic control method for precisely placing microcell plates (well-plate) on a grooved holder using tactile sensors, estimating the pose of the plate and holder through sliding while maintaining contact, achieving sub-millimeter level alignment.
Predicting Interaction Shape of Soft Continuum Robots using Deep Visual Models
Yunqi Huang, T. G. Thuruthel
Image TranslationRobotic IntelligenceConvolutional Neural NetworkImage
🎯 What it does: Proposed a deep visual model for predicting the shape of soft continuum robots during interaction with the environment.
Predicting Long-Term Human Behaviors in Discrete Representations via Physics-Guided Diffusion
Zhitian Zhang, Mo Chen
GenerationDiffusion modelAuto EncoderSequentialPhysics Related
🎯 What it does: Propose a long-term human trajectory prediction framework based on physics-guided diffusion models, which uses VQ-VAE to discretize continuous trajectories and contexts into a high-level latent action space, and predicts latent actions within this space.
Prediction of Acoustic Communication Performance for AUVs using Gaussian Process Classification
Yifei Gao, Daniel J. Stilwell
Robotic IntelligenceTabular
🎯 What it does: A probabilistic communication map based on the positions of transmitting and receiving AUVs is constructed using Gaussian process binary classification, incorporating vehicle position uncertainty into the mapping process, followed by experimental validation.
Predictive Coding for Decision Transformer
T. Luu, C. D. Yoo
TransformerReinforcement Learning
🎯 What it does: Proposed the Predictive Coding for Decision Transformer (PCDT) framework, improving the performance of decision transformers in offline goal-oriented reinforcement learning
Preliminary Result of Cury: A Backdrivable Leg Design Using Linear Actuators
Zhongtao Guan, Jiahao Chen
Robotic Intelligence
🎯 What it does: A robot leg prototype named Cury was designed, simulated, and experimentally studied, aiming to achieve minimal clearance and excellent reversible drive performance.
Preserving Relative Localization of FoV-Limited Drone Swarm via Active Mutual Observation
Lianjie Guo, Fei Gao
Robotic Intelligence
🎯 What it does: Proposes an active localization correction system that adjusts the camera direction in real-time via a yaw planner for relative state estimation in UAV formations with limited field of view;
Preventing Catastrophic Forgetting in Continuous Online Learning for Autonomous Driving
Rui Yang, Yassine Ruichek
Autonomous DrivingPoint Cloud
🎯 What it does: Proposed an online learning framework called Long-Short-Term Online Learning (LSTOL), aimed at preventing catastrophic forgetting and achieving long-term unsupervised learning in autonomous driving
Priority-Based Deadlock Recovery for Distributed Swarm Obstacle Avoidance in Cluttered Environments
Jiacheng He, Jinming Xu
OptimizationRobotic Intelligence
🎯 What it does: Propose a hierarchical priority mechanism for distributed drone swarms to achieve deadlock recovery in crowded environments through on-demand collision avoidance.
Privacy-Preserving Map-Free Exploration for Confirming the Absence of a Radioactive Source
Eric Lepowsky, Anirudha Majumdar
Safty and PrivacyRobotic IntelligencePhysics Related
🎯 What it does: Propose a robot verification program that realizes source verification tasks in mapless exploration, requiring neither nor leaking irrelevant site-specific information
Probabilistic Homotopy Optimization for Dynamic Motion Planning
Shayan Pardis, Sangbae Kim
OptimizationRobotic Intelligence
🎯 What it does: Propose a probabilistic homotopy optimization algorithm to solve optimization-based motion planning problems by searching through a sequence of optimization problems in a multi-dimensional homotopy parameter space, alternately performing solving and sampling stages, and using solutions from simpler problems as initial guesses for more complex problems.
Probabilistic Inference of Human Capabilities from Passive Observations
Peter Tisnikar, Matteo Leonetti
Robotic Intelligence
🎯 What it does: Inferring human capabilities through passive observation and utilizing the inferred capabilities to parameterize models, enabling personalized robot behavior during collaboration with humans; proposes the CAMO (Capability Modeling from Observations) model estimation algorithm.
Procedural generation of tunnel networks for unsupervised training and testing in underground applications
L. Cano, Alejandro R. Mosteo
Data SynthesisRobotic IntelligenceMeshGraph
🎯 What it does: Proposed a flexible tunnel network program generation method applicable to underground robot simulation.
Programming Passive Fingertip Deformation for Improved Grasping and Manipulation
Steffen Puhlmann, Hannes Höppner
Robotic Intelligence
🎯 What it does: Enhancing robotic grasping and manipulation performance through precise programming of passive deformation in soft fingers.
Progressive Query Refinement Framework for Bird's-Eye-View Semantic Segmentation from Surrounding Images
Dooseop Choi, KyoungWook Min
SegmentationAutonomous DrivingImage
🎯 What it does: Propose an advanced query refinement framework to apply multi-resolution residual learning and view transformation encoder in bird's-eye view semantic segmentation.
Progressive Representation Learning for Real-Time UAV Tracking
Changhong Fu, Jia Pan
Object TrackingRepresentation LearningVideo
🎯 What it does: Proposed an advanced representation learning framework called PRL-Track for real-time visual object tracking in unmanned aerial vehicles (UAVs).
Prompt-Driven Temporal Domain Adaptation for Nighttime UAV Tracking
Changhong Fu, Jia Pan
Object TrackingDomain AdaptationPrompt EngineeringGenerative Adversarial NetworkVideoBenchmark
🎯 What it does: Proposed a prompt-based temporal domain adaptation training framework for nighttime UAV tracking and implemented the TDA-Track tracker.
Proposal and Demonstration of a Robot Behavior Planning System Utilizing Video with Open Source Models in Real-World Environments
Yuki Akutsu, Koichi Osuka
GenerationRobotic IntelligenceDiffusion modelVideo
🎯 What it does: Developed and verified a video-based robot behavior planning system, demonstrating control for pick-and-place tasks on actual robots.
Proprioception Is All You Need: Terrain Classification for Boreal Forests
D. LaRocque, Franccois Pomerleau
ClassificationConvolutional Neural NetworkTime Series
🎯 What it does: Propose the BorealTC dataset and evaluate the performance of CNN and Mamba on terrain classification tasks
ProSIP: Probabilistic Surface Interaction Primitives for Learning of Robotic Cleaning of Edges
Christoph Unger, Andreas Kugi
Robotic Intelligence
🎯 What it does: Proposed a Probabilistic Surface Interaction Primitive (ProSIP) framework to learn human actions for cleaning edges on free-form surfaces, validated on an edge cleaning task on a bathroom sink.
PROSPECT: Precision Robot Spectroscopy Exploration and Characterization Tool
Nathaniel Hanson, T. Padır
Robotic IntelligenceMultimodalityPoint Cloud
🎯 What it does: Propose a sensorized end-effector and data acquisition strategy to capture the spectral features of objects and align them with 3D point clouds, generating a four-dimensional (position + spectral) model.
Proto-CLIP: Vision-Language Prototypical Network for Few-Shot Learning
P. JishnuJaykumar, Yu Xiang
ClassificationTransformerVision Language ModelContrastive LearningMultimodality
🎯 What it does: Proposes the Proto-CLIP framework, which constructs image and text prototypes using CLIP's image and text encoders, and performs few-shot learning through joint adaptation.
PS-Loc: Robust LiDAR Localization with Prior Structural Reference
Rui Li, Jingchuan Wang
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Developed a robust LiDAR localization framework based on prior structural references (e.g., floor plans), leveraging optimal transport matching methods and plane adjustment update strategies to enhance localization accuracy and robustness in dynamic scenes and rapid rotation conditions.
PS6D: Point Cloud Based Symmetry-Aware 6D Object Pose Estimation in Robot Bin-Picking
Yifan Yang, Jingtai Liu
Pose EstimationRobotic IntelligencePoint Cloud
🎯 What it does: Propose a pose estimation framework called PS6D based on point clouds, specifically designed for slender and symmetric industrial parts, utilizing multi-scale feature extraction, symmetry-aware rotation loss, and center-distance sensitive translation loss, along with two-stage clustering to achieve instance segmentation and pose regression.
Pseudo-Domain Adversarial Networks with Electrical Impedance Tomography for Electrode Offset Error
Gengchen Xu, Xiaojie Wang
Domain AdaptationGenerative Adversarial NetworkBiomedical Data
🎯 What it does: Proposes a transfer learning method called Pseudo-Domain Adversarial Network (PDAN) to address data errors caused by electrode displacement in electrical impedance tomography (EIT).
Pseudo-rigid body networks: learning interpretable deformable object dynamics from partial observations
Shamil Mamedov, Sebastian Trimpe
Explainability and InterpretabilityWorld ModelPhysics Related
🎯 What it does: Model DLO as a pseudo-rigid chain and jointly train with a dynamic network and physics-informed encoder to predict interpretable deformation dynamics from partial observations.
PSS-BA: LiDAR Bundle Adjustment with Progressive Spatial Smoothing
Jianping Li, Lihua Xie
OptimizationSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposed an LiDAR bundle adjustment method with advanced spatial smoothing to achieve high-quality point cloud construction
QO-Net: Query Optimization Underwater Object Detection Network
Jiandong Tian, Hongxin Xu
Object DetectionConvolutional Neural NetworkTransformerImage
🎯 What it does: Proposed a multi-scale feature enhancement and query optimization seawater depth detection network named QO-Net, and constructed a new seawater object detection dataset called UODD
QTrack: Embracing Quality Clues for Robust 3D Multi-Object Tracking
Jinrong Yang, Wenbing Tao
Object TrackingAutonomous DrivingMultimodality
🎯 What it does: Propose a novel 3D multi-object tracking framework QTrack, which utilizes quality learning to evaluate the confidence of position and velocity attributes, and adopts a quality-aware association strategy to achieve robust association.
Quadruped robot traversing 3D complex environments with limited perception
Yi Cheng, Bin Liang
Robotic Intelligence
🎯 What it does: Proposes an end-to-end learning-based quadruped robot motion controller that relies solely on its own proprioceptive sensors, capable of accurately detecting, localizing, and flexibly responding to collisions in unknown complex 3D environments;
Quaternion-Based Sliding Mode Control for Six Degrees of Freedom Flight Control of Quadrotors
Amin Yazdanshenas, Reza Faieghi
Robotic Intelligence
🎯 What it does: Proposed a six-degree-of-freedom sliding mode controller that achieves cascaded control of attitude and position.
QuerySOD: A Small Object Detection Algorithm Based on Sparse Convolutional Network and Query Mechanism
Zhengcai Cao, Meng Zhou
Object DetectionConvolutional Neural NetworkImage
🎯 What it does: Proposes the QuerySOD method for small object detection based on sparse convolutional networks and a query mechanism, constructing an extended feature pyramid network, sparse head, and query mechanism to leverage sparse feature maps;
QueSTMaps: Queryable Semantic Topological Maps for 3D Scene Understanding
Yash Mehan, Madhava Krishna
ClassificationSegmentationTransformerSimultaneous Localization and MappingText
🎯 What it does: Proposes a two-step process: first, using multi-channel occupancy representation to extract topological maps (floor plans) of indoor scenes, and then generating semantic labels for each room instance based on objects within rooms using CLIP-aligned features and self-attention transformers, while supporting natural language queries;
R2SNet: Scalable Domain Adaptation for Object Detection in Cloud–Based Robotic Ecosystems via Proposal Refinement
Michele Antonazzi, Nicola Basilico
Object DetectionDomain AdaptationRobotic Intelligence
🎯 What it does: Propose a downstream proposal refinement stage based on the lightweight DNN architecture R2SNet to assist cloud robots in domain adaptation for object detection in target environments locally.
RaceMOP: Mapless Online Path Planning for Multi-Agent Autonomous Racing using Residual Policy Learning
Raphael Trumpp, Marco Caccamo
Autonomous DrivingReinforcement Learning
🎯 What it does: Proposed a mapless online path planning method called RaceMOP, which utilizes residual policy learning to achieve high-speed overtaking decisions for multi-vehicle F1TENTH racing.
RADAR: Robotics Assembly by Demonstration via Augmented Reality
Wenhao Yang, Yunbo Zhang
Robotic IntelligenceBenchmark
🎯 What it does: Proposed and implemented the RADAR system, utilizing augmented reality (AR) technology to enable human-robot collaborative robotic assembly demonstrations.
Radiance Fields for Robotic Teleoperation
Maximum Wilder-Smith, Marco Hutter
Robotic IntelligenceNeural Radiance FieldGaussian SplattingImage
🎯 What it does: Achieved online training of radiance fields (NeRF, 3DGS) using multi-camera real-time data to replace traditional reconstruction visualization in a teleoperated robotic system
Rain-Reaper: Unmasking LiDAR-based Detector Vulnerabilities in Rain
Richard Capraru, B. Soong
Autonomous DrivingAdversarial AttackPoint Cloud
🎯 What it does: Designed a rainy-day attack method called Rain-Reaper based on genetic algorithms, which utilizes rainy conditions to identify critical detection points in 3D detectors and performs adversarial attacks.
Raising Body Ownership in End-to-End Visuomotor Policy Learning via Robot-Centric Pooling
Zheyu Zhuang, D. Kragic
Robotic IntelligenceContrastive LearningImage
🎯 What it does: Proposed a Robot-Centric Pooling (RcP) method to enhance the performance of end-to-end visual motion policies.
RAM-NAS: Resource-aware Multiobjective Neural Architecture Search Method for Robot Vision Tasks
Shouren Mao, Yongzhuo Gao
Knowledge DistillationRobotic IntelligenceNeural Architecture SearchImage
🎯 What it does: Proposed a resource-aware multi-objective neural architecture search method called RAM-NAS for robot vision tasks, which improves super-network pre-training while considering robot hardware resources. It employs mutual distillation between sub-networks and DKD loss for model distillation, and utilizes three types of robot edge hardware data to train a delay proxy, accelerating the search process.
RaNDT SLAM: Radar SLAM Based on Intensity-Augmented Normal Distributions Transform
Maximilian Hilger, B. Corves
Autonomous DrivingRobotic IntelligenceSimultaneous Localization and MappingMultimodalityPoint CloudBenchmark
🎯 What it does: Proposes the RaNDT SLAM framework, achieving fast and accurate robot trajectory estimation.
RATE: Real-time Asynchronous Feature Tracking with Event Cameras
Mikihiro Ikura, W. Stürzl
Object TrackingOptical Flow
🎯 What it does: Proposed a real-time asynchronous event camera feature tracking pipeline called RATE
RCAL:A Lightweight Road Cognition and Automated Labeling System for Autonomous Driving Scenarios
Jiancheng Chen, Changliang Xue
Autonomous Driving
🎯 What it does: Proposed and implemented a lightweight road cognition and automatic annotation system named RCAL, which utilizes lightweight road data collected from mass vehicles for road element vectorization and topological cognition. Multi-trip data is integrated on cloud servers to enhance accuracy and coverage. A joint node priority sampling strategy is proposed for balancing road scale and processing efficiency. Traffic flow information is further leveraged to improve the accuracy of road topological cognition.
React to This! How Humans Challenge Interactive Agents using Nonverbal Behaviors
Chuxuan Zhang, Angelica Lim
Robotic IntelligenceReinforcement Learning from Human FeedbackVideo
🎯 What it does: Collected and analyzed 1169 non-verbal interactions to study how humans challenge robots and virtual characters through facial and body behaviors.
Reactive Temporal Logic-based Planning and Control for Interactive Robotic Tasks
Farhad Nawaz, Nikolai Matni
Robotic Intelligence
🎯 What it does: Proposed a modular control architecture that integrates discrete task-level temporal logic planning with continuous dynamics system motion planning to generate safe and responsive interactive robot motion plans;
Real-time Bandwidth-efficient Occupancy Grid Map Synchronization for Multi-Robot Systems
Liuyu Shi, Fu Zhang
CompressionRobotic Intelligence
🎯 What it does: Propose a framework for synchronizing multi-robot occupancy grid maps (OGM) under low real-time communication bandwidth, designing a local OGM data structure and employing Hilbert space-filling curves for voxel sorting, significantly reducing communication data volume.
Real-time Bird’s-Eye-View Panoptic Segmentation for Monocular-based Indoor Navigation
Dawit Kim, Soon-Yyong Park
SegmentationData SynthesisComputational EfficiencyConvolutional Neural NetworkImage
🎯 What it does: Automatically generate bird's-eye view (BEV) training data for indoor environments using the physics engine of a simulator, and propose a lightweight network to achieve real-time BEV panoptic segmentation;
Real-time Coordinated Motion Generation: A Hierarchical Deep Predictive Learning Model for Bimanual Tasks
Genki Shikada, Tetsuya Ogata
Robotic IntelligenceRecurrent Neural Network
🎯 What it does: Propose a hierarchical deep predictive learning model for generating coordinated bimanual movements, with grasping experiments conducted on multiple objects using two robots
Real-time Coupled Centroidal Motion and Footstep Planning for Biped Robots
Tara Bartlett, Ian R. Manchester
OptimizationRobotic Intelligence
🎯 What it does: Proposed an algorithm that can quickly generate center of mass motion and footprint planning for SLIP-type bipedal robot models.
Real-time Dexterous Telemanipulation with an End-Effect-Oriented Learning-based Approach
Haoyang Wang, Lingfeng Tao
Robotic IntelligenceReinforcement Learning
🎯 What it does: Developed a learning framework called EFOLD based on end-effectors for real-time dexterous teleoperation, modeling the teleoperation process as a Markov game;
Real-time Hazard Prediction in Connected Autonomous Vehicles: A Digital Twin Approach
Sergio Barroso-Ramírez, Pedro Núñez Trujillo
Autonomous DrivingWorld Model
🎯 What it does: Propose a real-time hazard prediction model based on digital twin and validate it on actual autonomous electric vehicles
Real-Time Horizon Locking on Unmanned Surface Vehicles
Benjamin Kiefer, Andreas Zell
SegmentationAutonomous DrivingConvolutional Neural NetworkImage
🎯 What it does: Propose a real-time horizon locking method based on computer vision, utilizing real-time semantic segmentation to distinguish between sky, land, or water, achieving accurate horizon position locking.
Real-time Model Predictive Control with Zonotope-Based Neural Networks for Bipedal Social Navigation
Abdulaziz Shamsah, Ye Zhao
OptimizationRobotic Intelligence
🎯 What it does: Propose and verify a two-tier neural network framework based on zonotopes for predicting pedestrian trajectories and planning safe gaits for the bipedal robot Digit in crowded environments, while integrating this framework with a model predictive controller (MPC) to achieve real-time conflict avoidance and gait optimization.
Real-Time Particle Cluster Manipulation with Holographic Acoustic End-Effector under Microscope
Siyuan An, Song Liu
Robotic IntelligenceImageUltrasoundPhysics Related
🎯 What it does: A system was developed that utilizes an ultrasound phased array to achieve non-contact particle cluster manipulation under a microscope. By using a physics-driven deep learning algorithm, real-time phase holograms are computed to generate dynamic acoustic fields (optical-acoustic end effector), enabling the aggregation, rotation, and displacement of particle clusters.
Real-Time Path Generation and Alignment Control for Autonomous Curb Following
Yuanzhe Wang, Danwei Wang
Autonomous Driving
🎯 What it does: Propose a real-time path generation and alignment control scheme to support automatic lane-keeping.
Real-time Perceptive Motion Control using Control Barrier Functions with Analytical Smoothing for Six-Wheeled-Telescopic-Legged Robot Tachyon 3
Noriaki Takasugi, Yasunori Kawanami
OptimizationSafty and PrivacyRobotic Intelligence
🎯 What it does: A lightweight real-time perception motion control system is proposed, combining Control Barrier Functions (CBF) and SSAT to achieve safe real-time motion generation for the six-wheel extendable-legged robot Tachyon 3.
Real-time Robotic Flexible Needle Insertion In Deformable Living Organs Using Isolated Objective Constraint
Thuc Long Ha, H. Courtecuisse
OptimizationRobotic IntelligenceBiomedical Data
🎯 What it does: Propose a method for real-time control of robot fine needle insertion into deformable living organs, aiming to keep the needle's pivot point fixed on the skin while precisely navigating along a predefined trajectory despite organ displacement and deformation caused by respiration
Real-Time Semantic Segmentation in Natural Environments with SAM-assisted Sim-to-Real Domain Transfer
Han Wang, L. Teixeira
SegmentationDomain AdaptationComputational EfficiencyKnowledge DistillationConvolutional Neural NetworkTransformer
🎯 What it does: This paper proposes a two-step training pipeline for semantic segmentation in natural environments. First, a domain adaptation model is used for training, followed by pseudo label refinement using masks generated by the Segment Anything Model (SAM), ultimately distilling a MobileNetV3 model deployable in real-time.
Real-time terrain assessment and Bayesian-based path planning for off-road navigation
Tianwei Niu, Junzheng Wang
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: A framework for real-time terrain assessment and Bayesian path planning in unstructured, unknown environments is proposed.
Real-Time Truly-Coupled Lidar-Inertial Motion Correction and Spatiotemporal Dynamic Object Detection
C. Gentil, Teresa Vidal-Calleja
Object DetectionAutonomous DrivingOptimizationSimultaneous Localization and MappingPoint CloudTime Series
🎯 What it does: Proposes a real-time tightly coupled motion distortion correction method by integrating LiDAR with inertial measurement units (IMU), and achieves map-free dynamic object detection based on this method.
Realistic Rainy Weather Simulation for LiDARs in CARLA Simulator
Donglin Yang, Xinyu Cai
Data SynthesisAutonomous DrivingPoint Cloud
🎯 What it does: Implement LiDAR point cloud simulation under rainy conditions, including raindrop spray and light intensity effects, using the CARLA simulator to generate synthetic rainy weather data for data augmentation.
Reality Fusion: Robust Real-time Immersive Mobile Robot Teleoperation with Volumetric Visual Data Fusion
Ke Li, Frank Steinicke
Robotic IntelligenceGaussian SplattingPoint Cloud
🎯 What it does: Proposes Reality Fusion, a robotic teleoperation system that integrates local depth sensors with 3D Gaussian splats (3DGS) rendering, enabling egocentric and exocentric remote operation in immersive VR.
Reconfigurable Multi-Rotor for High-Precision Physical Interaction
Joshua Taylor, Efe Camci
Robotic Intelligence
🎯 What it does: Designed a drone equipped with a tilted rotor and a reconfigurable landing gear, which can transform into a front-mounted dual-finger gripper for high-precision physical interaction with vertical cylindrical targets (e.g., trees).
Reconfigurable Robot Identification from Motion Data
Yuhang Hu, Hod Lipson
Pose EstimationRobotic IntelligenceMeta LearningTime SeriesSequential
🎯 What it does: Achieve self-modeling by inferring robot pose configuration through body perception
Reconfigurable Soft Gripper Based on Eversion and Electroadhesion for Cluttered Environments
Dana Ragab, Hareesh Godaba
Robotic Intelligence
🎯 What it does: Propose a soft gripper with two steerable, length-adjustable reversible fingers, and design a multi-layer single-insulation layer electro-adhesive pad that can be safely integrated into the reversible fingers.
Recover: A Neuro-Symbolic Framework for Failure Detection and Recovery
Cristina Cornelio, Mohammed Diab
Anomaly DetectionRobotic IntelligenceTransformerLarge Language Model
🎯 What it does: Proposes Recover, a neuro-symbolic framework for online failure detection and recovery in robot task execution, and introduces the OntoThor ontology in the AI2Thor simulated kitchen environment.
Recovering Missed Detections in an Elevator Button Segmentation Task
Nicholas Verzic, Justin W. Hart
Object DetectionSegmentationConvolutional Neural NetworkImage
🎯 What it does: Propose a segmentation and detection method specifically for recovering missed elevator buttons and labels by the initial segmentation model, and construct a corresponding dataset.
Recurrent Non-Rigid Point Cloud Registration
Yue Cao, Hongdong Li
Pose EstimationRecurrent Neural NetworkPoint CloudBenchmark
🎯 What it does: Proposed a non-rigid point cloud registration framework based on a recursive architecture
Redefining Data Pairing for Motion Retargeting Leveraging a Human Body Prior
Xiyana Figuera, Hyemin Ahn
Pose EstimationRobotic Intelligence
🎯 What it does: Propose MR.HuBo, which collects high-quality upper limb robot-human paired data by randomly sampling robot poses and converting them into human poses, and introduces human body priors to filter extreme poses; simultaneously propose a two-stage supervised motion retargeting network.
Reducing Cognitive Load in Teleoperating Swarms of Robots through a Data-Driven Shared Control Approach
Enrico Turco, T. L. Baldi
Robotic Intelligence
🎯 What it does: Propose a data-driven shared control method that enables a single operator to effectively control nine degrees of freedom in a multi-robot formation, with experimental validation conducted in a simulated narrow cylindrical path environment.
Reducing Performance Variability and Overcoming Limited Spatial Ability: Targeted Training for Remote Robot Teleoperation
Tsung-Chi Lin, Chien-Ming Huang
Robotic IntelligenceDrug Discovery
🎯 What it does: Proposed and verified a targeted training method aimed at enabling users with varying abilities to achieve consistent proficiency in remote robotic teleoperation.
REF2-NeRF: Reflection and Refraction aware Neural Radiance Field
Wooseok Kim, Takeshi Oishi
GenerationNeural Radiance FieldImage
🎯 What it does: Proposed a NeRF-based glass case scene modeling method that can simultaneously handle reflection and refraction phenomena.
Refining Airway Segmentation Through Breakage Filling and Leakage Reduction Using Point Clouds
Yan Hu, Yang Song
SegmentationConvolutional Neural NetworkPoint CloudBiomedical Data
🎯 What it does: Propose a refined segmentation method for pulmonary airways based on point cloud.
Refractive COLMAP: Refractive Structure-from-Motion Revisited
M. She (University of Kiel), Kevin Köser
Data SynthesisPose EstimationDepth EstimationSimultaneous Localization and MappingImagePhysics Related
🎯 What it does: Proposes a complete refractive structure from motion (RSfM) framework for underwater 3D reconstruction using refractive cameras (flat and dome).
Reinforcement Learning Control for Autonomous Hydraulic Material Handling Machines with Underactuated Tools
Filippo A. Spinelli, Marco Hutter
Robotic IntelligenceReinforcement Learning
🎯 What it does: Developed a reinforcement learning-based controller capable of simultaneously commanding cabin joints and robotic arms for precise and safe control of autonomous hydraulic material handling machines.
Reinforcement Learning for Active Search and Grasp in Clutter
T. Pitcher, Jen Jen Chung
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed an active search strategy that balances camera movement and occlusion removal in cluttered environments to search for and retrieve target objects.
Reinforcement Learning of Dolly-In Filming Using a Ground-Based Robot
Philip Lorimer, Wenbin Li
Robotic IntelligenceReinforcement Learning
🎯 What it does: Using reinforcement learning to automate close-up push-pull shots for free-roaming ground shooting robots
Reinforcement Learning with Generalizable Gaussian Splatting
Jiaxu Wang, Renjing Xu
Robotic IntelligenceReinforcement LearningGaussian Splatting
🎯 What it does: Proposes a general Gaussian splatting framework GSRL as an environment representation for reinforcement learning tasks, validated in the RoboMimic environment.
RelationGrasp: Object-Oriented Prompt Learning for Simultaneously Grasp Detection and Manipulation Relationship in Open Vocabulary
Songting Liu, Haiyue Zhu
Object DetectionPose EstimationRobotic IntelligenceTransformerPrompt EngineeringImage
🎯 What it does: Propose the RelationGrasp unified framework, which employs a Transformer encoder-decoder architecture to simultaneously accomplish open-vocabulary object detection, manipulation relation reasoning, and grasp pose detection.
ReLoc-Aligner : Orientation-aware Scene Descriptor for Re-Localization within a 3D Point Cloud Map
SungJoon Cho, Jun-Sik Kim
Pose EstimationPoint Cloud
🎯 What it does: Proposed an orientation-aware scene descriptor called ReLoc-Aligner for relocalization of 3D point clouds.