ICRA 2024 Papers — Page 14
IEEE International Conference on Robotics and Automation · 1760 papers
Reinforcement Learning in a Safety-Embedded MDP with Trajectory Optimization
Fan Yang, David Held
OptimizationSafty and PrivacyRobotic IntelligenceReinforcement Learning
🎯 What it does: Propose a safe reinforcement learning method that embeds safety constraints into a modified MDP action space and converts RL-generated action sequences into safe trajectories through trajectory optimization.
Reinforcement Learning with Human Feedback for Realistic Traffic Simulation
Yulong Cao, Marco Pavone
Reinforcement Learning from Human FeedbackReinforcement Learning
🎯 What it does: Developed a framework called TrafficRLHF based on RLHF to enhance the realism of existing traffic models.
Relaxed Hover Solution Based Control for a Bi-copter with Rotor and Servo Stuck Failure
Haixin Zhao, Quan Quan
OptimizationRobotic Intelligence
🎯 What it does: Studied the control of twin-rotor helicopters under servo or rotor lock failures, proposed a relaxed hover solution, and designed an LQR-based attitude reduction controller and a cascaded PID position controller, validated through numerical simulations.
RELEAD: Resilient Localization with Enhanced LiDAR Odometry in Adverse Environments
Zhiqiang Chen, Ming Liu
Autonomous DrivingOptimizationSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Propose the RELEAD method to address the degradation of LiDAR scan matching in adverse environments;
ReLU-QP: A GPU-Accelerated Quadratic Programming Solver for Model-Predictive Control
Arun L. Bishop, Zachary Manchester
Optimization
🎯 What it does: Proposed ReLU-QP, a GPU-accelerated quadratic programming solver that can solve high-dimensional control problems in real-time.
Remote Control of Untethered Magnetic Robots within a Lumen using X-Ray-Guided Robotic Platform
Leendert-Jan W. Ligtenberg, Islam S. M. Khalil
Robotic IntelligenceBiomedical Data
🎯 What it does: Demonstrates the remote operation of untethered magnetic robots within vascular cavities using an X-ray-guided robotic platform.
RenderOcc: Vision-Centric 3D Occupancy Prediction with 2D Rendering Supervision
Mingjie Pan, Shanghang Zhang
Depth EstimationAutonomous DrivingNeural Radiance FieldImage
🎯 What it does: Proposes the RenderOcc framework, which employs a NeRF-style 3D volume representation and volumetric rendering techniques to train a 3D occupancy prediction model using only 2D semantic and depth labels, and introduces an auxiliary ray method to address sparse viewpoint issues.
ReorientDiff: Diffusion Model based Reorientation for Object Manipulation
Utkarsh Aashu Mishra, Yongxin Chen
Pose EstimationRobotic IntelligenceVision-Language-Action ModelDiffusion modelImage
🎯 What it does: Proposed a ReorientDiff method based on diffusion models, using visual inputs and target language prompts to plan intermediate reorientation poses during robotic manipulation;
Representing On-Orbit Rendezvous and Proximity Operations with Fully-Actuated Multirotor Aerial Platforms
Alessandro Garzelli, A. Ollero
Robotic IntelligenceSimultaneous Localization and MappingPhysics Related
🎯 What it does: Recreate a zero-gravity environment indoors using a full-actuated multirotor and verify spacecraft proximity operation control algorithms by simulating free flight/float conditions;
Representing Robot Geometry as Distance Fields: Applications to Whole-body Manipulation
Yiming Li, S. Calinon
OptimizationRobotic Intelligence
🎯 What it does: Propose a method to represent robot geometry as a distance field (RDF), encoding signed distances precisely on each link using Bernstein polynomials and generating differentiable SDFs in joint space for whole-body control and collision avoidance.
Resampling-free Particle Filters in High-dimensions
Akhilan Boopathy, I. Fiete
Pose EstimationVideo
🎯 What it does: Proposed a resampling-free particle filter to alleviate the particle depletion problem in high-dimensional state spaces
Research on bionic foldable wing for flapping wing micro air vehicle
Shengjie Xiao, Xilun Ding
Robotic Intelligence
🎯 What it does: Designed and experimentally verified a foldable wing inspired by the hind wings of beetles, established a folding motion model, and implemented folding and unfolding using shape memory alloys and torsion springs, tested its lift and attitude torque, and validated that its performance is comparable to an optimized non-foldable wing.
Residual-NeRF: Learning Residual NeRFs for Transparent Object Manipulation
B. Duisterhof, Jeffrey Ichnowski
Depth EstimationConvolutional Neural NetworkNeural Radiance FieldImage
🎯 What it does: Propose the Residual-NeRF method, which first learns the background NeRF and trains the residual NeRF and Mixnet to improve depth perception and training speed for transparent objects.
Resilient Legged Local Navigation: Learning to Traverse with Compromised Perception End-to-End
Jin Jin, Marco Hutter
Robotic IntelligenceReinforcement LearningWorld Model
🎯 What it does: Train a local navigation strategy based on reinforcement learning (RL) that can reconstruct environmental information and respond accordingly when perceptual distortion occurs, while fusing joint motion perception with external sensory inputs.
Resolving Loop Closure Confusion in Repetitive Environments for Visual SLAM through AI Foundation Models Assistance
Hongzhou Li, Guang Tan
TransformerVision Language ModelSimultaneous Localization and MappingImage
🎯 What it does: Proposes a loop closure detection method that utilizes pre-trained AI foundation models to extract semantic anchors (e.g., door numbers), addressing perceptual confusion in visual SLAM within repetitive environments.
Rethinking Imitation-based Planners for Autonomous Driving
Jie Cheng, Ming Liu
Autonomous DrivingBenchmark
🎯 What it does: Conduct a comprehensive study using the nuPlan platform to explore the core features of self-planning, effective data augmentation techniques, reveal the imitation learning gap, and propose the PlanTF baseline model.
Rethinking Social Robot Navigation: Leveraging the Best of Two Worlds
Amir Hossain Raj, Xuesu Xiao
Robotic IntelligencePoint Cloud
🎯 What it does: Developed a hybrid planner that switches between geometric navigation and deep learning planning to enhance social compliance
RETOM: Leveraging Maneuverability for Reactive Tool Manipulation using Wrench-Fields
Felix Eberle, Sami Haddadin
Robotic Intelligence
🎯 What it does: Proposed a vector field-based real-time reactive planner that achieves effective manipulation of tools in complex environments by considering tool geometry features and rotational torque, and enhances performance through encoding using robot capability metrics and mass distribution methods.
RETRO: Reactive Trajectory Optimization for Real-Time Robot Motion Planning in Dynamic Environments
Apan Dastider, Mingjie Lin
OptimizationRobotic Intelligence
🎯 What it does: Proposes the RETRO framework for real-time generation of robot trajectories in dynamic environments.
RGB-based Category-level Object Pose Estimation via Decoupled Metric Scale Recovery
Jiaxin Wei, Pan Ji
Pose EstimationImage
🎯 What it does: Propose an RGB method that decouples 6D pose estimation and size estimation, leveraging a monocular pre-trained estimator to extract local geometric information and restoring the true scale through category-level scale statistics, finally using RANSAC-PnP for robust pose solving.
RGBD-based Image Goal Navigation with Pose Drift: A Topo-metric Graph based Approach
Shuhao Ye, Yue Wang
Pose EstimationAutonomous DrivingReinforcement LearningSimultaneous Localization and MappingImage
🎯 What it does: Proposed a drift-compensated topological-metric graph for environment mapping and localization, and enhanced navigation efficiency through a subgoal selection strategy based on reinforcement learning.
RGBManip: Monocular Image-based Robotic Manipulation through Active Object Pose Estimation
Boshi An, Hao Dong
Pose EstimationRobotic IntelligenceReinforcement LearningImage
🎯 What it does: Utilizing a monocular camera (eye-in-hand), active multi-view perception is achieved on the parallel gripper of a robotic arm, and this information is used to estimate the target's 6D pose to complete the robotic grasping.
RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shot
Haoshu Fang, Cewu Lu
Robotic IntelligenceImageVideoTextMultimodalityBenchmarkAudio
🎯 What it does: Collected over 110,000+ multimodal robot operation sequences, each equipped with visual, force, audio data, action information, as well as corresponding human demonstration videos and language descriptions.
RIC: Rotate-Inpaint-Complete for Generalizable Scene Reconstruction
Isaac Kasahara, Volkan Isler
RestorationGenerationDepth EstimationTransformerLarge Language ModelVision Language ModelImageMesh
🎯 What it does: First, inpainting is performed on scene images rendered from different perspectives using a large vision-language model, and then the completed image is elevated to a full 3D structure and texture by predicting normal vectors and solving for depth.
Ricmonk: A Three-Link Brachiation Robot with Passive Grippers for Energy-Efficient Brachiation
Shourie S. Grama, Frank Kirchner
OptimizationRobotic Intelligence
🎯 What it does: Designed, analyzed, and evaluated a three-link catenary robot named RicMonk equipped with a passive hook-shaped gripper
RIDE: Self-Supervised Learning of Rotation-Equivariant Keypoint Detection and Invariant Description for Endoscopy
M. Karaoglu, A. Ladikos
Pose EstimationConvolutional Neural NetworkBiomedical Data
🎯 What it does: Proposed and implemented a learning-based, rotation equivariant keypoint detection and invariant description method called RIDE, which is self-supervised trained to address rotation issues in endoscopic images.
RIDER: Reinforcement-Based Inferred Dynamics via Emulating Rehearsals for Robot Navigation in Unstructured Environments
S. Siva, Maggie B. Wigness
Robotic IntelligenceReinforcement LearningWorld Model
🎯 What it does: Proposes a framework called RIDER that utilizes reinforcement learning to learn robot-terrain interaction dynamics in a latent space, enabling autonomous navigation.
RISeg: Robot Interactive Object Segmentation via Body Frame-Invariant Features
Howard H. Qian, Kaiyu Hang
SegmentationRobotic IntelligenceConvolutional Neural NetworkContrastive LearningOptical FlowImagePoint Cloud
🎯 What it does: Proposes an interactive perception workflow that corrects unseen object instance segmentation errors in static images through robot interaction and body-frame invariant features.
Risk-aware Control for Robots with Non-Gaussian Belief Spaces
M. Vahs, Jana Tumova
Robotic Intelligence
🎯 What it does: For risk-aware control of robots in non-Gaussian Bayesian spaces, Bayesian states and dynamics are defined, a safe set is constructed, and a controller is designed to remain within the safe set, ensuring bounded safety violation risks.
Risk-aware Trajectory Prediction by Incorporating Spatio-temporal Traffic Interaction Analysis
Divya Thuremella, L. Kunze
Autonomous DrivingTime SeriesSequential
🎯 What it does: By analyzing location information and speed related to high-risk interactions in the dataset, and adopting a position- and speed-based re-weighting technique during training to enhance trajectory prediction in high-risk scenarios.
Risk-Bounded Online Team Interventions via Theory of Mind
Yuening Zhang, Brian C. Williams
Reinforcement Learning
🎯 What it does: Propose a risk-bounded AI team assistant that integrates cognitive planning with POMDP, intervening only when the team's failure probability exceeds a threshold or potential deadlocks occur.
Risk-Inspired Aerial Active Exploration for Enhancing Autonomous Driving of UGV in Unknown Off-Road Environments
Rongchuan Wang, Wenjie Song
Autonomous DrivingOptimizationPoint Cloud
🎯 What it does: Propose a risk-based UAV active exploration system that guides unmanned ground vehicles (UGV) to safely navigate unknown off-road environments through UAV's field of view and flexibility;
Risk-Predictive Planning for Off-Road Autonomy
L. L. Beyer, S. Karaman
Autonomous DrivingOptimization
🎯 What it does: Propose an algorithm for risk prediction planning in an offline environment
Risk-Sensitive Extended Kalman Filter
Armand Jordana, L. Righetti
OptimizationRobotic Intelligence
🎯 What it does: Proposed a risk-sensitive extended Kalman filter (Risk-Sensitive EKF) that adjusts estimates according to control objectives, enabling safe output feedback model predictive control (MPC).
RiskBench: A Scenario-based Benchmark for Risk Identification
Chi-Hsi Kung, Yi-Ting Chen
Safty and PrivacyBenchmark
🎯 What it does: Introduced RiskBench, a scenario-based large-scale risk identification benchmark; designed a scenario taxonomy and augmentation pipeline to systematically collect real-world risks across diverse scenarios; evaluated ten algorithms' capabilities in risk detection and localization, risk prediction, and decision support; conducted extensive experiments and summarized future research directions.
Road Obstacle Detection based on Unknown Objectness Scores
Chihiro Noguchi, Masao Yamanaka
Object DetectionAnomaly DetectionAutonomous DrivingConvolutional Neural Network
🎯 What it does: Combining pixel-level anomaly detection with object detection methods, using a semantic segmentation network with a sigmoid head to provide pixel-level anomaly scores and objectness scores, and proposing an unknown objectness score for road obstacle detection.
RoboAgent: Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunking
Homanga Bharadhwaj, Vikash Kumar
Robotic IntelligenceAgentic AIMultimodality
🎯 What it does: Proposes an efficient framework MT-ACT for training general-purpose robotic agents capable of performing multi-task operations, leveraging semantic enhancement and action representation to generate a single strategy RoboAgent with 12 unique skills, demonstrating generalization capabilities across daily tasks in 38 kitchen scenarios.
RoboBall: An All-Terrain Spherical Robot with a Pressurized Shell
M. Oevermann, Robert O. Ambrose
Robotic Intelligence
🎯 What it does: Designed and manufactured a soft inflatable spherical shell with an internal 2-DOF pendulum-driven all-terrain spherical robot, and proposed a model for pendulum control and shell dynamics, subsequently verifying its performance through experimental validation and field tests.
RoboGrind: Intuitive and Interactive Surface Treatment with Industrial Robots
Benjamin Alt, Marco F. Huber
Robotic Intelligence
🎯 What it does: Proposed the RoboGrind system, achieving intuitive and interactive automation for industrial robots in surface processing tasks (e.g., grinding, sanding, polishing).
RoboHop: Segment-based Topological Map Representation for Open-World Visual Navigation
Sourav Garg, Ian Reid
Robotic IntelligenceGraph Neural NetworkVision-Language-Action ModelImageText
🎯 What it does: Propose a topology map representation based on segments, constructing a graph with segments as nodes, connecting segments from adjacent images, and using graph convolution to update segment-level descriptors.
RoboKeyGen: Robot Pose and Joint Angles Estimation via Diffusion-based 3D Keypoint Generation
Yang Tian, Hao Dong
Pose EstimationRobotic IntelligenceDiffusion model
🎯 What it does: Proposes a framework that decomposes robot pose and joint angle estimation into 2D keypoint detection and 3D keypoint lifting, leveraging diffusion models to achieve conditional 3D keypoint generation for efficient pose and joint angle estimation.
RoboLLM: Robotic Vision Tasks Grounded on Multimodal Large Language Models
Zijun Long, Gerardo Aragon Camarasa
RecognitionObject DetectionSegmentationRobotic IntelligenceTransformerLarge Language ModelSupervised Fine-TuningVision-Language-Action ModelMultimodality
🎯 What it does: Propose the RoboLLM framework, which leverages the BEiT-3 multimodal large language model to uniformly address tasks such as object detection, segmentation, and recognition in robot vision.
Robot Fine-Tuning Made Easy: Pre-Training Rewards and Policies for Autonomous Real-World Reinforcement Learning
Jingyun Yang, Chelsea Finn
Robotic IntelligenceReinforcement LearningVision Language Model
🎯 What it does: Proposed a no-reset reinforcement learning fine-tuning system called RoboFuME, which leverages pre-trained multi-task control policies and online self-improvement to achieve target task learning with minimal human intervention.
Robot Interaction Behavior Generation based on Social Motion Forecasting for Human-Robot Interaction
Esteve Valls Mascaro, Dongheui Lee
GenerationRobotic IntelligenceTransformer
🎯 What it does: Propose social motion prediction in a shared human-robot representation space and synthesize robot motions that interact with humans based on this.
Robot Navigation in Unseen Environments using Coarse Maps
Chengguang Xu, Lawson L. S. Wong
Robotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Proposes the Coarse Map Navigator (CMN) framework, which utilizes a rough map to enable robot navigation in unknown environments, addressing issues of new visual observations and map errors/misalignment.
Robot Synesthesia: In-Hand Manipulation with Visuotactile Sensing
Ying Yuan, Xiaolong Wang
Robotic IntelligenceReinforcement LearningMultimodalityPoint Cloud
🎯 What it does: Achieve in-hand manipulation by combining visual and tactile perception, proposing the Robot Synesthesia point cloud tactile representation.
Robot Task Planning Under Local Observability
Max Merlin, G. Konidaris
Robotic IntelligenceSimultaneous Localization and MappingWorld Model
🎯 What it does: Propose a locally observable Markov decision process (LO-MDP) and design a three-phase planning workflow, integrating existing Markov planners, achieving success in mobile robot tasks.
Robot Trajectron: Trajectory Prediction-based Shared Control for Robot Manipulation
Pinhao Song (KU Leuven), R. Detry (KU Leuven)
Robotic Intelligence
🎯 What it does: Predict the arm's movement trajectory and use the predictor to implement shared control, reducing the operator's cognitive load.
Robot-Assisted Navigation for Visually Impaired through Adaptive Impedance and Path Planning
P. Balatti, Arash Ajoudani
Robotic IntelligencePoint Cloud
🎯 What it does: A framework based on a mobile manipulator for robot-assisted navigation is proposed, helping visually impaired individuals move safely in unfamiliar environments through a physically coupled arm. The framework includes a mobile base, manipulator, obstacle avoidance unit, human leg tracking algorithm, and adaptive pulling planner;
Robot-Dependent Traversability Estimation for Outdoor Environments using Deep Multimodal Variational Autoencoders
Matthias Eder, Gerald Steinbauer-Wagner
Robotic IntelligenceAuto EncoderMultimodality
🎯 What it does: Studied a cross-robot off-road traversability estimation method using Deep Multimodal Variational Autoencoders (DMVAEs).
RoboTAP: Tracking Arbitrary Points for Few-Shot Visual Imitation
Mel Vecerík, Jonathan Scholz
Robotic IntelligenceMeta Learning
🎯 What it does: Propose a dense tracking method based on the Track-Any-Point (TAP) model for rapidly learning robot visual imitation from a small amount of demonstration data;
Robotic capillary insertion to the Xenopus oocyte using microscopic image analysis and QCR force sensor
Kazusa Otani, Fumihito Arai
Robotic IntelligenceImage
🎯 What it does: Developed a 3D oocyte manipulation system based on a stereo microscope, utilizing sequential calibration methods for spatial positioning, and combined optical detection with quartz crystal resonator (QCR) force sensors to achieve automatic capillary insertion in TEVC experiments.
Robotic Constrained Imitation Learning for the Peg Transfer Task in Fundamentals of Laparoscopic Surgery
Kento Kawaharazuka, Masayuki Inaba
Robotic IntelligenceImage
🎯 What it does: Achieve the peg transfer task in FLS using imitation learning, and propose an implementation strategy that achieves more accurate imitation learning using only monocular images.
Robotic Craniomaxillofacial Osteotomy System Using Acoustic 3D Registration *
Jiayu Zhu, Junchen Wang
Robotic IntelligenceBiomedical DataUltrasound
🎯 What it does: Propose a non-invasive image-to-patient registration method based on handheld ultrasound 3D reconstruction, and develop a craniofacial bone resection robotic system capable of safe human interaction and trajectory planning.
Robotic Exploration through Semantic Topometric Mapping
Scott Fredriksson, G. Nikolakopoulos
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Propose a strategy for robot exploration utilizing semantic topological mapping.
Robotic Grasping of Harvested Tomato Trusses Using Vision and Online Learning
Luuk van den Bent, R. Babuška
Robotic IntelligenceConvolutional Neural NetworkImagePoint CloudAgriculture Related
🎯 What it does: A deep learning visual system is proposed for identifying and grasping picked tomato stems in a loading box, combined with an online learning grasping pose ranking algorithm to achieve grasping without tactile sensors and geometric models.
Robotic Manipulation of Hand Tools: The Case of Screwdriving
Ling Tang, Yuechuan Xue
OptimizationRobotic Intelligence
🎯 What it does: Studied the grasping and force control methods for robotic arm/hand coordination using a hand tool (screwdriver) for screw rotation, proposed an inverse chain-based force control scheme, and validated it in simulations and Shadow Hand experiments.
Robotic Mosaic Atomic Force Microscopy Through Sequential Imaging and Multiview Iterative Closest Points Method
Freddy Romero Leiro (ISIR laboratory), Mokrane Boudaoud (ISIR laboratory)
Robotic IntelligenceImagePhysics Related
🎯 What it does: Developed an AFM-in-SEM robotic system that achieves long-range, drift-eliminated AFM topography mapping by fusing multiple AFM topography maps using the GPA-ICP algorithm.
Robotic Offline RL from Internet Videos via Value-Function Learning
Chethan Bhateja, Aviral Kumar
Representation LearningRobotic IntelligenceReinforcement LearningVideo
🎯 What it does: Proposed and implemented the V-PTR system, which leverages large-scale internet video data for pre-training the value function, providing superior feature representations for robot offline reinforcement learning.
RobotPerf: An Open-Source, Vendor-Agnostic, Benchmarking Suite for Evaluating Robotics Computing System Performance
V'ictor Mayoral-Vilches, V. Reddi
Robotic IntelligenceBenchmark
🎯 What it does: Developed an open-source, vendor-independent benchmark suite called RobotPerf to evaluate the computing performance of ROS 2-based robots on different hardware platforms, covering complete robot workflows and providing both black-box and gray-box testing methods.
RoboVQA: Multimodal Long-Horizon Reasoning for Robotics
P. Sermanet, Yuan Cao
Data-Centric LearningRobotic IntelligenceTransformerVision Language ModelVideoTextMultimodality
🎯 What it does: Proposed an scalable, bottom-up, intrinsically diverse data collection scheme using multiple agents (robots, humans, and humans with grasping tools) to collect real-world video-text pairs with user requests in three office buildings, and released the RoboVQA dataset containing 29,520 unique instructions and 829,502 (video, text) pairs; simultaneously trained a video-conditioned model RoboVQA-VideoCoCa for executing high-level vision-based reasoning tasks and guiding real robots to complete long-horizon tasks.
Robust 3D Object Detection from LiDAR-Radar Point Clouds via Cross-Modal Feature Augmentation
Jianning Deng, C. X. Lu
Object DetectionAutonomous DrivingMultimodalityPoint Cloud
🎯 What it does: Propose a robust 3D object detection framework based on cross-modal feature enhancement, which can achieve high-precision detection under single-modal input.
Robust and Dexterous Dual-arm Tele-Cooperation using Adaptable Impedance Control
Keyhan Kouhkiloui Babarahmati, S. Vijayakumar
Robotic Intelligence
🎯 What it does: Proposed and verified an adaptive impedance control method for dual-arm remote collaboration and teleoperation; experimental results verified its robustness and flexibility.
Robust and Energy-Efficient Control for Multi-task Aerial Manipulation with Automatic Arm-switching
Ying Wu, Hui Cheng
OptimizationRobotic Intelligence
🎯 What it does: Propose a learning-based control algorithm to achieve automatic arm switching in multi-task aerial operations, supporting online trajectory optimization and tracking. A deep neural network is used to classify and learn torques generated by different arms and their interaction with the environment, enabling energy-efficient trajectory planning.
Robust and Remote Center of Cyclic Motion Control for Redundant Robots with Partially Unknown Structure
Long Jin, Mei Liu
Robotic Intelligence
🎯 What it does: Developed an acceleration-level cyclic motion remote center (ARC2M) control scheme, and proposed a computational method for estimating unknown end-effector parameters under noise influence.
Robust Balancing Control of Biped Robots for External Forces
H. Park, Jung Hoon Kim
Robotic Intelligence
🎯 What it does: A controller synthesis method was developed to ensure the acceptable upper limit of external forces for bipedal robots at a given level.
Robust Co-Design of Canonical Underactuated Systems for Increased Certifiable Stability
Federico Girlanda, Frank Kirchner
Optimization
🎯 What it does: Proposes a robust co-design algorithm RTC-D to enhance the verifiable stability of underactuated systems.
Robust Collaborative Perception against Temporal Information Disturbance
Xunjie He, Yufeng Yue
Object DetectionAutonomous DrivingTransformerVideo
🎯 What it does: Proposed a robust collaborative perception framework that predicts perception information under information perturbations
Robust Collaborative Perception without External Localization and Clock Devices
Zixing Lei, Yanfeng Wang
Pose EstimationAutonomous DrivingGraph Neural Network
🎯 What it does: Proposes a robust collaborative perception system that does not rely on external localization and clock devices, achieving spatiotemporal alignment through the FreeAlign module.
Robust Control for Bidirectional Thrust Quadrotors under Instantaneously Drastic Disturbances
Zujian Chen, Hui Cheng
Robotic Intelligence
🎯 What it does: Propose a robust control framework for bi-directional thrust quadrotors against instantaneous severe disturbances
Robust In-Hand Manipulation with Extrinsic Contacts
Boyuan Liang, Devesh K. Jha
Robotic Intelligence
🎯 What it does: Propose a robust grasp-in-hand manipulation method that can move objects and adjust internal hand posture while maintaining the external contact mode.
Robust Indoor Localization with Ranging-IMU Fusion
Fan Jiang, Jing Dong
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Achieving indoor positioning by fusing low-power wireless ranging with low-cost IMU inertial measurements
Robust MITL planning under uncertain navigation times
Alexis Linard, Jana Tumova
OptimizationRobotic IntelligenceReinforcement Learning
🎯 What it does: The study addresses task planning in environments with uncertain navigation durations using Metric Interval Temporal Logic (MITL), and proposes a method to maximize the temporal robustness of robotic tasks through a combination of Mixed Integer Linear Programming (MILP) and recursive horizon planning.
Robust Model Predictive Control with Control Barrier Functions for Autonomous Surface Vessels
Wei Wang, Daniela Rus
Autonomous DrivingOptimization
🎯 What it does: Designed and verified a robust model predictive control based on control barrier functions for trajectory tracking of autonomous surface vessels, which can maintain the system within a predefined safe area under external disturbances.
Robust Multi-Robot Global Localization with Unknown Initial Pose based on Neighbor Constraints
Yaojie Zhang, Wei Feng
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Proposed a data association algorithm based on neighbor constraints to enhance the robustness of multi-robot global localization.
Robust Quadrupedal Locomotion via Risk-Averse Policy Learning
Jiyuan Shi, Xuelong Li
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed a robust locomotion method for quadruped robots based on a risk-sensitive perspective, utilizing distributed value functions and risk-preference policy learning to enhance performance in uncertain environments.
Robust Surgical Tool Tracking with Pixel-based Probabilities for Projected Geometric Primitives
Christopher D'Ambrosia, Michael C. Yip
Object TrackingRobotic IntelligenceImageBiomedical Data
🎯 What it does: Estimate the coordinate transformation between the robot and the camera, as well as the measurement error of the surgical tool's joint angles, and achieve tool tracking through image-based insertion axis detection algorithms and probabilistic models.
Robustified Time-optimal Collision-free Motion Planning for Autonomous Mobile Robots under Disturbance Conditions
Shuhao Zhang, Jan Swevers
OptimizationRobotic Intelligence
🎯 What it does: A robust time-optimal collision avoidance motion planning method for autonomous mobile robots (AMR) affected by random process noise and measurement noise is proposed, enabling navigation from an initial state to a goal state without colliding with obstacles.
Robustifying a Policy in Multi-Agent RL with Diverse Cooperative Behaviors and Adversarial Style Sampling for Assistive Tasks
Takayuki Osa, Tatsuya Harada
Robotic IntelligenceReinforcement LearningBenchmark
🎯 What it does: Proposed a framework that enhances the robustness of caregiver strategies in multi-agent reinforcement learning by training diverse caregiver responses and adopting adversarial sampling, evaluated on the Assistive Gym task.
RoCo: Dialectic Multi-Robot Collaboration with Large Language Models
Zhao Mandi, Shuran Song
Robotic IntelligenceLarge Language ModelPrompt EngineeringTextBenchmark
🎯 What it does: Proposes a multi-robot collaboration method leveraging large language models (LLMs), utilizing LLMs for high-level communication and low-level path planning, and achieving interpretable and flexible collaboration through dialogue.
Rolling with Planar Parametric Curves for Real-time Robot Locomotion Algorithms
Adwait Mane, Christian Hubicki
OptimizationRobotic Intelligence
🎯 What it does: Proposed and derived a closed-form dynamic model for the rolling motion of two planar smooth curves, applying it to the sagittal plane walking problem, including simulating non-slipping rolling on terrain with variable curvature and generating control signals to achieve stability.
RoSSO: A High-Performance Python Package for Robotic Surveillance Strategy Optimization Using JAX
Yohan J. John, F. Bullo
OptimizationRobotic Intelligence
🎯 What it does: Proposed the RoSSO Python package for computing efficient random patrol routes for single or multi-robot teams, solving Markov chain optimization problems.
RTS-GT: Robotic Total Stations Ground Truthing dataset
Maxime Vaidis, François Pomerleau
Data SynthesisRobotic IntelligenceTime SeriesBenchmark
🎯 What it does: Propose the RTS-GT dataset and use three robotic total stations (RTS) to track a mobile platform, generating 6-DOF ground truth trajectories
RUMP: Robust Underwater Motion Planning in Dynamic Environments of Fast-moving Obstacles
H. B. Amundsen, Eleni Kelasidi
OptimizationRobotic Intelligence
🎯 What it does: Proposed the RUMP framework, achieving robust real-time underwater motion planning in dynamic obstacle environments using path optimization and real-time nonlinear solvers;
S2R-ViT for Multi-Agent Cooperative Perception: Bridging the Gap from Simulation to Reality
Jinlong Li, Hongkai Yu
Domain AdaptationAutonomous DrivingTransformerGenerative Adversarial NetworkPoint Cloud
🎯 What it does: Proposed and studied the S2R-ViT, a multi-agent collaborative perception transfer learning framework from simulation to reality, addressing deployment gap and feature gap.
Safe and Individualized Motion Planning for Upper-limb Exoskeleton Robots Using Human Demonstration and Interactive Learning
Yu Chen, Xiang Li
Robotic IntelligenceReinforcement Learning from Human FeedbackTime SeriesBiomedical Data
🎯 What it does: A motion planning scheme for upper limb exoskeleton robots based on human demonstration and interactive learning is proposed. First, reference trajectories are generated by learning from healthy subjects, then personalized trajectories are iteratively optimized using patient interaction data, and finally, assistance is achieved by tracking under a variable damping model.
Safe Deep Policy Adaptation
Wenli Xiao, Guanya Shi
Safty and PrivacyRobotic IntelligenceSupervised Fine-TuningReinforcement LearningBenchmark
🎯 What it does: Propose the SafeDPA framework, which jointly learns adaptive policies and dynamic models in simulation, fine-tunes with a small amount of real-world data, and combines a safety filter based on control barrier functions to achieve safe deep policy adaptation.
Safe Execution of Learned Orientation Skills with Conic Control Barrier Functions
Zheng Shen, Sami Haddadin
OptimizationSafty and PrivacyOrdinary Differential Equation
🎯 What it does: Propose a method to safely execute learned directional skills within a constrained region, using stable DS on SO(3), extracting time-varying conical constraints from expert demonstrations, and constraining the evolution of DS through Conic Control Barrier Function (CCBF).
Safe Multi-Robot Exploration using Symbolic Control
Manas Sashank Juvvi, Pushpak Jagtap
Safty and PrivacyRobotic Intelligence
🎯 What it does: Proposes a modular multi-robot safety exploration framework that identifies safe frontier goals while considering the dynamics of each robot system, and achieves collision avoidance among unknown obstacles and between robots through symbolic control to ensure task completion.
Safe POMDP Online Planning via Shielding
Shili Sheng, Lu Feng
Safty and PrivacyReinforcement Learning
🎯 What it does: Propose to achieve safe POMDP online planning through shields, compute and integrate shields that satisfy almost-sure reach-avoid specifications, integrate them into the POMCP algorithm, and propose four shield methods.
Safe Receding Horizon Motion Planning with Infinitesimal Update Interval
Inkyu Jang, H. J. Kim
Autonomous DrivingOptimization
🎯 What it does: The study examines the behavior of safe recursive visibility motion planning as the update interval approaches infinity, reformulating the trajectory optimization problem into a time derivative form to obtain a quadratic programming problem that directly outputs safe inputs.
Safe Table Tennis Swing Stroke with Low-Cost Hardware
Francesco Cursi, Jianye Hao
OptimizationRobotic Intelligence
🎯 What it does: Proposes a safety motion planning framework that generates robotic arm trajectories satisfying environmental safety constraints and scaling to joint motion limits, enabling ping-pong hitting at hardware limits.
Safe-By-Design Digital Twins for Human-Robot Interaction: A Use Case for Humanoid Service Robots
Jon Skerlj, Sami Haddadin
🎯 What it does: Proposed a secure digital twin framework that integrates a real-time safety module to limit the base and arm speeds of humanoid service mobile robots during human interaction, ensuring compliance with biomechanical injury thresholds and verifying the effectiveness of this method in simulated human environments.
Safety Optimized Reinforcement Learning via Multi-Objective Policy Optimization
Homayoun Honari, Homayoun Najjaran
Reinforcement Learning
🎯 What it does: Proposed a model-free safe reinforcement learning algorithm (SORL) based on multi-objective policy optimization, simultaneously optimizing optimality and safety;
Safety Verification of Closed-loop Control System with Anytime Perception
Lipsy Gupta, Pavithra Prabhakar
🎯 What it does: Propose a safety analysis framework for closed-loop control systems, and provide a general procedure using reachable set computation; for classical discrete-time linear systems and variable update rate extensions, precise polyhedral computation methods and over-approximation techniques are respectively provided.
Safety-Conscious Pushing on Diverse Oriented Surfaces with Underactuated Aerial Vehicles
Tong Hui, M. Fumagalli
Safty and PrivacyRobotic Intelligence
🎯 What it does: Established a safety assessment process for drone pushing tasks on surfaces with various orientations, predicting the saturation levels of each actuator and planning safe experiments based on the assessment results.
Safety-critical Control of Quadrupedal Robots with Rolling Arms for Autonomous Inspection of Complex Environments
Jaemin Lee, A. Ames
Robotic Intelligence
🎯 What it does: Proposed a safety-critical control framework for quadruped robots equipped with wheel arms to enable autonomous patrolling in complex multi-layer environments, achieving smooth transitions between columns and footpad re-planning.
Safety-Critical Coordination of Legged Robots via Layered Controllers and Forward Reachable Set based Control Barrier Functions
Jeeseop Kim, A. Ames
Robotic Intelligence
🎯 What it does: Propose a hierarchical controller based on control barrier functions and forward reachable sets to achieve safe collaboration and trajectory tracking for legged robots in dynamic environments.
Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing
Haolan Liu, Jishen Zhao
Autonomous DrivingSafty and PrivacyReinforcement LearningAuto Encoder
🎯 What it does: Generate safety-critical scenarios using deep reinforcement learning through sequence editing (adding agents or modifying trajectories).
SAGE-ICP: Semantic Information-Assisted ICP
Jiaming Cui, Liang Li
Pose EstimationAutonomous DrivingConvolutional Neural NetworkSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Integrating semantic information to improve LiDAR point-to-point ICP for pose estimation