IROS 2024 Papers — Page 2
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers
Abstraction of the Body Ability of the Transformer Robot System for the Transportation and Installation of Heavy Objects in Land and Underwater Environments
Tasuku Makabe, Masayuki Inaba
Robotic IntelligenceGraph
🎯 What it does: This paper proposes abstracting the physical capabilities of a deformable robot system, designed for heavy object transportation and installation tasks in terrestrial and underwater environments, into states detectable by sensors to achieve adaptive planning in response to environmental and target changes.
Accelerating Model Predictive Control for Legged Robots through Distributed Optimization
Lorenzo Amatucci, Claudio Semini
OptimizationRobotic Intelligence
🎯 What it does: By decomposing robot dynamics into parallel subsystems and utilizing ADMM to achieve consistency between subsystems, the performance of model predictive control for legged robots is improved.
Accurate and Efficient Loop Closure Detection With Deep Binary Image Descriptor and Augmented Point Cloud Registration
Jialiang Wang, Benwen Chen
Simultaneous Localization and MappingImagePoint Cloud
🎯 What it does: Proposed a loop closure detection method based on camera-LiDAR fusion
Accurate power consumption estimation method makes walking robots energy efficient and quiet
Giorgio Valsecchi, Marco Hutter
OptimizationRobotic IntelligenceRecurrent Neural NetworkReinforcement Learning
🎯 What it does: Proposed an LSTM-based precise actuator power consumption estimation method, integrated into the Isaac Gym framework for training RL policies that minimize energy consumption.
Accurately Tracking Relative Positions of Moving Trackers based on UWB Ranging and Inertial Sensing without Anchors
Rayan Armani, Christian Holz
OptimizationSimultaneous Localization and MappingTime SeriesSequential
🎯 What it does: Proposes a relative positioning system relying solely on mobile tracking nodes, utilizing UWB ranging and 9-DOF magnetic/inertial sensors without requiring fixed base stations.
Active Human Pose Estimation via an Autonomous UAV Agent
Jingxi Chen, Y. Aloimonos
Pose EstimationConvolutional Neural NetworkNeural Radiance Field
🎯 What it does: This paper proposes an active 2D human pose estimation system for autonomous drones, which enhances pose estimation accuracy by repositioning the camera to obtain a better view.
Active Information Gathering for Long-Horizon Navigation Under Uncertainty by Learning the Value of Information
R. I. Arnob, Gregory J. Stein
Autonomous DrivingGraph Neural NetworkReinforcement Learning
🎯 What it does: Proposed a planning strategy based on value information, utilizing graph neural networks to predict the benefits of exploration actions, thereby achieving active information collection navigation in long-term partially mapped environments.
Active Learning for Forward/Inverse Kinematics of Redundantly-driven Flexible Tensegrity Manipulator
Yuhei Yoshimitsu, Shuhei Ikemoto
Robotic IntelligenceAuto Encoder
🎯 What it does: Proposes an active learning framework for flexible redundant-driven multi-degree-of-freedom tension structure manipulators, used for forward and inverse kinematic modeling, incorporating VAE networks and cross-entropy method for sample selection.
Active Learning-augmented Intention-aware Obstacle Avoidance of Autonomous Surface Vehicles in High-traffic Waters
Mingi Jeong, Alberto Quattrini Li
Autonomous DrivingOptimizationRecurrent Neural Network
🎯 What it does: Proposes an active learning enhanced, intent-aware obstacle avoidance method to improve the safe navigation of ASVs in high-traffic waters.
Active Loop Closure for OSM-guided Robotic Mapping in Large-Scale Urban Environments
Wei Gao, Hui Kong
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Proposes an active closed-loop loop detection mechanism, enabling the robot to actively replan previously GPS-based trajectories using OSM, revisit previous locations to trigger backend optimization, and reduce pose estimation errors and uncertainty.
Active Neural Mapping at Scale
Zi-Feng Kuang, Hongbin Zha
Neural Radiance FieldSimultaneous Localization and Mapping
🎯 What it does: Proposed an active mapping system based on NeRF that enables efficient and robust exploration in large-scale indoor environments.
Active Pose Refinement for Textureless Shiny Objects using the Structured Light Camera
Jun Yang, Steven L. Waslander
Pose EstimationDepth EstimationOptimizationPoint Cloud
🎯 What it does: Proposes a 6D pose optimization and next-best-view prediction framework using a structured light camera for fine-grained pose refinement of textureless and glossy objects.
Active propulsion noise shaping for multi-rotor aircraft localization
Gabriele Serussi, Alexander M. Bronstein
Robotic IntelligenceAudio
🎯 What it does: Proactively control and shape the propulsion noise generated by multirotor rotor blades, use neural networks for self-noise localization, and achieve precise and robust positioning through learning time-varying rotor phase modulation.
Active Scout: Multi-Target Tracking Using Neural Radiance Fields in Dense Urban Environments
Christopher D. Hsu, Pratik Chaudhari
Object TrackingNeural Radiance FieldImage
🎯 What it does: Build a city's NeRF representation online using RGB and depth images, compute information gain from this representation to actively explore unknown areas and track multiple ground targets, demonstrating that within 300 steps, it can localize 20 static targets in a custom simulator using Philadelphia and New York City OSM data.
Active Semantic Mapping and Pose Graph Spectral Analysis for Robot Exploration
Rongge Zhang, Giovanni Beltrame
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Proposes a novel active metric semantic SLAM method that treats exploration as an active semantic mapping problem, utilizing semantic mutual information and pose graph connectivity to evaluate trajectories and select optimal strategies.
Active Vehicle Re-localization Based on Non-repetitive LiDAR with Gimbal Motion Strategy
Xin Wu, Ming Yang
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Using a single non-repetitive scanning LiDAR mounted on a dual-axis rotating gimbal to achieve active relocalization of the vehicle;
ActiveRIR: Active Audio-Visual Exploration for Acoustic Environment Modeling
Arjun Somayazulu, Kristen Grauman
Robotic IntelligenceReinforcement LearningImageMultimodalityAudio
🎯 What it does: Studied a mobile agent based on active audio-visual exploration, utilizing reinforcement learning strategies for active acoustic sampling to efficiently construct acoustic models and occupancy maps of indoor environments.
ActNeRF: Uncertainty-aware Active Learning of NeRF-based Object Models for Robot Manipulators using Visual and Re-orientation Actions
Saptarshi Dasgupta, Rohan Paul
Pose EstimationDepth EstimationRobotic IntelligenceNeural Radiance Field
🎯 What it does: Using a NeRF-based model, the robot actively learns the complete 3D model of objects through visual and relocalization actions, and rapidly constructs internal representations under unknown poses.
Adapting Skills to Novel Grasps: A Self-Supervised Approach
Georgios Papagiannis, Edward Johns
Pose EstimationRobotic IntelligenceImage
🎯 What it does: The study presents a method for adapting operation trajectories defined under a single grasp pose to new grasp poses
Adaptive Control Barrier Functions for Near-Structure ROV Operations
Malte von Benzon, Simon Pedersen
Safty and PrivacyRobotic Intelligence
🎯 What it does: Proposed a safety controller based on adaptive control barrier functions (CBF) specifically designed to enhance the operational safety and efficiency of autonomous remotely operated vehicles (ROVs) in inspection, maintenance, and repair (IMR) tasks.
Adaptive Feedforward Super-Twisting Sliding Mode Control of Parallel Kinematic Manipulators With Real-Time Experiments
Hussein Saied, Clovis Francis
Robotic Intelligence
🎯 What it does: Propose an adaptive feedforward super-twisting sliding mode control algorithm to address the tracking control problem of parallel manipulators, and conduct real-time experiments on a 3-DOF Delta robot.
Adaptive Model Predictive Control for Differential-Algebraic Systems towards a Higher Path Accuracy for Physically Coupled Robots
Xin Ye, Sören Hohmann
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Investigate the path tracking accuracy problem of physically coupled robots, propose a differential algebraic system model, estimate uncertain kinematic parameters online, and design an adaptive model predictive controller to achieve robot cooperative control.
Adaptive multi-altitude search and sampling of sparsely distributed natural phenomena
Jessica E. Todd, Dana R. Yoerger
OptimizationImagePoint Cloud
🎯 What it does: Proposes a Sparse Adaptive Search and Sample (SASS) algorithm to autonomously locate sparsely distributed targets in unknown underwater environments.
Adaptive Passivation of Admittance Controllers by Bypassing Power to Null Space on Redundant Manipulators
Yeoil Yun, J. Koo
Robotic Intelligence
🎯 What it does: Propose a passivity-preserving gain adjustment method applicable to redundant robots, achieving adaptive hierarchical control by projecting non-passive power into the null space while maintaining the robot's response to human inputs;
Adaptive Planning with Generative Models under Uncertainty
Pascal Jutras-Dubé, Aniket Bera
GenerationRobotic IntelligenceReinforcement LearningWorld ModelSequential
🎯 What it does: Proposed an adaptive planning strategy based on a generative model that can execute multiple consecutive actions without immediate replanning
Adaptive Smith Predictor Fractional Control of a Tele-operated Flexible Link Robot *
Saddam Gharab, V. F. Batlle
Robotic Intelligence
🎯 What it does: Control of a remote manipulator with flexible links using an adaptive Smith predictor, utilizing motor angle encoders and base strain gauges for measurement, and employing a fractional-order controller for closed-loop end-effector position control.
Adaptive Social Force Window Planner with Reinforcement Learning
Mauro Martini, Luis Merino
Autonomous DrivingReinforcement Learning
🎯 What it does: Proposed an adaptive social planner that dynamically adjusts the weight parameters in the trajectory evaluation cost function using deep reinforcement learning, integrating the classic dynamic window approach (DWA) with the social force model (Social Force Model);
Adaptive Splitting of Reusable Temporal Monitors for Rare Traffic Violations
Craig Innes, S. Ramamoorthy
Autonomous Driving
🎯 What it does: Proposes a method that combines rare event sampling techniques with online specification monitoring, using adaptive multilevel splitting to decompose simulations into partial trajectories, and evaluates the distance of these trajectories to failure through signal temporal logic (STL) robustness metrics, while caching partial robustness values for reuse across multiple sampling stages to efficiently estimate the probability of safety violations in autonomous vehicles during simulations.
Adaptive Stochastic Nonlinear Model Predictive Control with Look-ahead Deep Reinforcement Learning for Autonomous Vehicle Motion Control
Baha Zarrouki, Johannes Betz
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: Proposes an adaptive stochastic nonlinear model predictive control method based on deep reinforcement learning for autonomous vehicle motion control
Adaptive Trajectory Database Learning for Nonlinear Control with Hybrid Gradient Optimization
Kuan-Yu Tseng, G. Dullerud
OptimizationReinforcement Learning
🎯 What it does: Developed an experience-based learning method EHGO, which utilizes a trajectory database optimized under a reference dynamics and adaptively improves control policies on real systems through hybrid gradient optimization (GRILC);
Adaptive Visual-Aided 4D Radar Odometry Through Transformer-Based Feature Fusion
Yuanfan Zhang, Jie Liu
Autonomous DrivingTransformerSimultaneous Localization and MappingImagePoint Cloud
🎯 What it does: Propose an unsupervised deep learning trajectory estimation method based on the fusion of visual and 4D radar data to address odometry problems under different weather conditions.
Adv3D: Generating 3D Adversarial Examples for 3D Object Detection in Driving Scenarios with NeRF
Leheng Li, Yingke Chen
Object DetectionAutonomous DrivingAdversarial AttackNeural Radiance Field
🎯 What it does: This paper proposes using NeRF to generate 3D adversarial examples for attacking 3D object detection in driving scenarios;
Advanced Handheld Micro-Surgical System using an Hall Sensor and a Magnet Trocar for Retinal Microsurgery
Myung Ho Lee, Cheol Song
Robotic IntelligenceBiomedical Data
🎯 What it does: Proposed a handheld minimally invasive surgical tool with a 1-DOF mechanism and a 3-axis Hall sensor to reduce hand tremors during retinal surgery.
Advancements in Radar Odometry
Matteo Frosi, Matteo Matteucci
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Building upon existing state-of-the-art radar odometry methods, this paper proposes improvements including filtering, motion compensation, surface point calculation, Gaussian smoothing, one-to-many radar scan registration, and pose refinement, while incorporating ICP alignment to enhance local scene understanding.
AdvDiffuser: Generating Adversarial Safety-Critical Driving Scenarios via Guided Diffusion
Yuting Xie, Long Chen
Autonomous DrivingDiffusion modelImageVideo
🎯 What it does: Proposes AdvDiffuser, an adversarial framework that generates safety-critical driving scenarios through guided diffusion;
Adversarial Attack on Trajectory Prediction for Autonomous Vehicles with Generative Adversarial Networks
Jiping Fan, Guoqiang Li
Autonomous DrivingAdversarial AttackRecurrent Neural NetworkGenerative Adversarial NetworkTime SeriesSequential
🎯 What it does: Proposed a GAN-based trajectory prediction adversarial attack method and developed an LSTM-based Adv-GAN model.
AEGO: Modeling Attention for HRI in Ego-Sphere Neural Networks
H. F. Chame, Rachid Alami
Representation LearningRobotic IntelligenceMultimodality
🎯 What it does: Propose the AEGO architecture for modeling attention in human-computer interaction, fusing multimodal information and integrating them in a shared representation space.
Ag2Manip: Learning Novel Manipulation Skills with Agent-Agnostic Visual and Action Representations
Puhao Li, Siyuan Huang
Robotic IntelligenceVision-Language-Action ModelVideo
🎯 What it does: Proposes the Ag2Manip framework, which leverages agent-agnostic visual and action representations to learn new robotic manipulation tasks.
Agent-Agnostic Centralized Training for Decentralized Multi-Agent Cooperative Driving
Shengchao Yan, Wolfram Burgard
Autonomous DrivingReinforcement Learning
🎯 What it does: Proposed an asymmetric actor-critic model that utilizes single-agent reinforcement learning to learn decentralized collaborative driving strategies.
Agile and Safe Trajectory Planning for Quadruped Navigation with Motion Anisotropy Awareness
Wentao Zhang, Lijun Zhu
OptimizationRobotic Intelligence
🎯 What it does: Proposes a navigation framework for quadruped robots that considers the anisotropy of their motion, encompassing kinodynamic trajectory generation, nonlinear trajectory optimization, and nonlinear model predictive control.
AGL-Net: Aerial-Ground Cross-Modal Global Localization with Varying Scales
Tianrui Guan, Dinesh Manocha
Autonomous DrivingSimultaneous Localization and MappingImagePoint Cloud
🎯 What it does: Proposed AGL-Net, a learning-based global localization method that performs two-stage matching using LiDAR point clouds and satellite maps, achieving scale-invariant feature representation through skeleton features and scale alignment;
Agonist-Antagonist Pouch Motors: Bidirectional Soft Actuators Enhanced by Thermally Responsive Peltier Elements
Trevor Exley, Amir Jafari
Robotic Intelligence
🎯 What it does: Developed a Mylar-based pneumatic motor that utilizes the reversible thermal effect of the Peltier junction to achieve muscle expansion and contraction;
AirCrab: A Hybrid Aerial-Ground Manipulator with An Active Wheel
Muqing Cao, Lihua Xie
Robotic Intelligence
🎯 What it does: Developed a hybrid aerial-ground robot named AirCrab equipped with a single-actuator wheel and a 3-degree-of-freedom robotic arm, achieving functions of reducing position drift and improving control accuracy through ground contact.
AirShot: Efficient Few-Shot Detection for Autonomous Exploration
Zihan Wang, Sebastian Scherer
Object DetectionMeta LearningImage
🎯 What it does: Propose the AirShot framework to achieve few-shot object detection.
Aligning Learning with Communication in Shared Autonomy
Joshua Hoegerman, Dylan P. Losey
Robotic Intelligence
🎯 What it does: The study investigates the impact of conveying auxiliary information learned by robots to human operators in shared autonomy, develops experimental and theoretical models, conducts online and offline user studies, modifies existing robot learning algorithms to better interpret human input with communication signals, and evaluates their effectiveness in user learning and control interaction.
All-day Depth Completion
Vadim Ezhov, Alex Wong
Depth EstimationImagePoint Cloud
🎯 What it does: Propose a method for depth estimation under different lighting conditions during the day and night, utilizing multi-sensor fusion—projecting synchronous sparse laser point clouds onto the image plane to generate sparse depth maps, along with camera images as input;
Alternative Connection Radius for Asymptotic Optimality in RRT*
Rahul Shome
Autonomous DrivingOptimization
🎯 What it does: Propose a connection radius method based on a minimized RRT* variant, demonstrating the asymptotic optimality of RRT* while eliminating large parameters that are difficult to estimate.
AMCO: Adaptive Multimodal Coupling of Vision and Proprioception for Quadruped Robot Navigation in Outdoor Environments
Mohamed Bashir Elnoor, Dinesh Manocha
SegmentationRobotic IntelligenceMultimodality
🎯 What it does: Proposes an adaptive multi-modal coupling method for visual and proprioceptive navigation in quadruped robots, combining three cost maps (general knowledge map, passage history map, and current proprioceptive map) into a fused passage cost map, which is then used to plan gait and speed.
An Active and Dexterous Bionic Torso for a Quadruped Robot*
Ruyue Li, Mengnan Zhou
Robotic Intelligence
🎯 What it does: Design and manufacture an active bionic torso to simulate the multi-directional motion of a quadruped's spine, and integrate it into a quadruped robot to verify its feasibility in real-world environments.
An Actor-Critic Reinforcement Learning Scheme for Reactive 3D Optimal Motion Planning Based on Fluid Dynamics
Marios Malliaropoulos, Kostas J. Kyriakopoulos
Autonomous DrivingOptimizationReinforcement LearningPhysics Related
🎯 What it does: Propose a new method for solving 3D optimal motion planning using fluid dynamics potential streamlines, employing model-based Actor-Critic reinforcement learning to approximate the HJB equation, generating closed-loop smooth and natural navigation solutions;
An Adaptive Robotic Exoskeleton for Comprehensive Force-Controlled Hand Rehabilitation
Nikolas J. Wilhelm, Rainer H. H. Burgkart
Robotic IntelligenceBiomedical Data
🎯 What it does: Developed and validated an adaptive hand robotic exoskeleton for force control rehabilitation in CRPS patients.
An Agile Robotic Penguin Driven by Submersible Geared Servomotors: Various Maneuvers by Active Feathering of the Wings
Taiki Shimooka, Hiroto Tanaka
Robotic Intelligence
🎯 What it does: Designed and tested an agile underwater robot penguin equipped with dual-degree-of-freedom wings, achieving rapid acceleration, emergency stopping, rolling, pitching, and yawing maneuvers through feathering adjustment of the wings;
An Attention-aware Deep Reinforcement Learning Framework for UAV-UGV Collaborative Route Planning
Md Safwan Mondal, Pranav A. Bhounsule
Autonomous DrivingOptimizationTransformerReinforcement Learning
🎯 What it does: Designed and implemented an attention mechanism-based deep reinforcement learning (DRL) framework for cooperative path planning of drones and unmanned ground vehicles (UGVs), capable of sequentially selecting actions to construct UAV and UGV routes and determine charging meeting points.
An Autonomous, 3D Printed, Waterjet-Powered, Open-Source Robotic Trimaran for Environmental Inspection and Monitoring
Reuben O'Brien, Minas V. Liarokapis
Autonomous DrivingRobotic Intelligence
🎯 What it does: Proposed an autonomous, 3D-printed, jet-propelled open-source trimaran platform for environmental monitoring and inspection
An Efficient Coverage Method for Irregularly Shaped Terrains
Yuxuan Tang, Lei Chen
Optimization
🎯 What it does: A grid-based model is proposed, redefining coverage path optimization as searching for the maximum Hamiltonian subgraph in a grid graph, and introducing a Hamiltonian cycle expansion strategy along with a low-repetition coverage path planner to quickly generate efficient full-coverage paths.
An Efficient Position Reconfiguration Approach for Maximizing Lifetime of Fixed-wing Swarm Drones
Han Liu, Kai Huang
Optimization
🎯 What it does: Propose an efficient fixed-wing UAV swarm position reconfiguration method aimed at maximizing flight duration.
An Ejecting System for Autonomous Takeoff of Flapping-Wing Robots
Xu Jiang, Aiguo Song
OptimizationRobotic Intelligence
🎯 What it does: Designed and implemented a ground-based launch system that utilizes the symmetric slider-crank mechanism (S-SCM) to convert spring energy storage into takeoff velocity for the dive-wing robot (FWR), achieving autonomous takeoff.
An Intelligent Robotic System for Perceptive Pancake Batter Stirring and Precise Pouring
Xinyuan Luo, Wenzhen Yuan
Robotic Intelligence
🎯 What it does: Designed and implemented a smart cooking robot capable of autonomously stirring batter, estimating batter properties, and precisely pouring into specified shapes through tactile sensing and control algorithms.
An LSTM-based Model to Recognize Driving Style and Predict Acceleration
Jiaxing Lu, He Bai
RecognitionAutonomous DrivingRecurrent Neural NetworkTime SeriesSequential
🎯 What it does: Developed an LSTM-based model for identifying driving styles and predicting vehicle acceleration, and evaluated its performance using a simulation test bench with lane-changing experiments involving five vehicles.
An MR Safe Double-Arch Needle Insertion Robot with Scissor-Folding Mechanism for Abdominal Percutaneous Interventions*
Ziting Liang, A. Stilli
Robotic IntelligenceMagnetic Resonance Imaging
🎯 What it does: Designed and evaluated an MR-safe 5-degree-of-freedom (DOF) parallel table-top dual-arch needle insertion robot equipped with a scissor-folding mechanism for abdominal interventions.
An Observability Constrained Downward-Facing Optical-Flow-Aided Visual-Inertial Odometry
Dan Liu, Shuo Li
Pose EstimationSimultaneous Localization and MappingOptical FlowImage
🎯 What it does: Propose a tightly-coupled estimator that integrates measurements from a downward-looking optical flow sensor into the visual-inertial odometry (VIO) framework to enhance localization accuracy
An Octopus-Inspired-Configuration Sensor Array Concept toward Torso-Oriented Magnetic Localization Task and Simulation Verification
Yichong Sun, Zheng Li
Optimization
🎯 What it does: Proposed a magnetic sensor array configuration based on the flexibility of octopus tentacles and their covering morphology, conducted geometric analysis under a constant curvature model, developed a magnetic tracking optimization algorithm, and validated its effectiveness in a simulation environment.
An Online Automatic Calibration Method for Infrastructure-Based LiDAR-Camera via Cross-modal Object Matching
Tao Wang, Ming Yang
Autonomous DrivingOptimizationConvolutional Neural NetworkImageMultimodalityPoint Cloud
🎯 What it does: Propose an infrastructure LiDAR-camera online automatic calibration method based on cross-modal target registration, eliminating the need for manual targets and initial pose guessing.
An Online Rcm Adjusting System for Robot-Assisted Retinal Surgeries
Jun Xia, Kai Huang
Robotic IntelligenceBiomedical Data
🎯 What it does: Proposes an online RCM (remote center of motion) adjustment strategy that dynamically adjusts the remote center point between the surgical instrument and trocar to limit external forces.
An Optical Interferometer-based Force Sensor System for Enhancing Precision in Epidural Injection Procedure
Gichan Cho, Cheol Song
Optical Flow
🎯 What it does: Integrate an optical interferometer force sensor into a commercial epidural needle, calibrate the system to establish a correspondence between system output and actual force, develop a graphical interface to identify the puncture point based on sudden force drops, and subsequently evaluate the puncture depth and success rate through user studies.
An Optimization-Based Planner with B-spline Parameterized Continuous-Time Reference Signals
Chuyuan Tao, N. Hovakimyan
OptimizationComputational Efficiency
🎯 What it does: Proposed a B-spline parameterized optimization planner (BSPOP) that generates continuous-time control inputs, addressing the frequency gap between the planner and controller, and operates under limited onboard computational resources.
An Optimization-based Scheme for Real-time Transfer of Human Arm Motion to Robot Arm
Zhelin Yang, Sami Haddadin
OptimizationRobotic Intelligence
🎯 What it does: Proposed an optimization-based real-time humanoid motion transfer framework that transfers human arm movements in real-time to robotic arms.
An Origami-Inspired Pneumatic Continuum Module with Active Variable Stiffness
Zhuowen Li, Hesheng Wang
Robotic Intelligence
🎯 What it does: Proposed a pneumatic continuum module with high contraction ratio, bidirectional driving, and active variable stiffness, and completed its design, manufacturing, modeling, and performance verification.
An Ultrafast Multi-object Zooming System Based on Low-latency Stereo Correspondence
Qing Li, I. Ishii
Computational EfficiencyOptical FlowVideo
🎯 What it does: Developed an ultra-fast multi-target zoom system using a panoramic high-frame-rate (HFR) stereo camera and an oscillator-based reflective pan-tilt-zoom (PTZ) camera. By leveraging high-speed motion information for stereo matching, the system maps 3D position data to PTZ control voltage, achieving time-division multiplexed high-speed, clear image capture for multiple targets.
Analysis of Lockable Passive Prismatic and Revolute Joints
Abdur Rosyid, B. El-Khasawneh
Robotic IntelligencePhysics RelatedOrdinary Differential Equation
🎯 What it does: A systematic analysis of stress, positional error, and friction in locked passive prismatic and rotary joints
Anchor-Oriented Localized Voronoi Partitioning for GPS-denied Multi-Robot Coverage
Aiman Munir, Ramviyas Parasuraman
Robotic Intelligence
🎯 What it does: Propose the anchor-oriented coverage (AOC) method, generating dynamic local Voronoi partitions based on shared anchors, and design a consensus coordination algorithm to enable robots to reach agreement on the coverage workspace within a relative coordinate framework.
Answerability Fields: Answerable Location Estimation via Diffusion Models
Daich Azuma, M. Kawanabe
RecognitionDiffusion modelMultimodalityPoint Cloud
🎯 What it does: Proposes Answerability Fields (AnsFields) to predict the feasibility of question-answering at different locations in indoor environments, inferring AnsFields from top-down scene views and questions using diffusion models, and subsequently employing these predicted AnsFields for 3D question-answering.
AnytimeFusion: Parameter-free RGB Camera-Radar Sensor Fusion Algorithm in Complex Maritime Situations
Yeongha Shin, Jinwhan Kim
Autonomous DrivingOptimizationMultimodality
🎯 What it does: Proposes a parameter-free RGB camera-radar sensor fusion algorithm called AnytimeFusion for obstacle localization in complex maritime environments.
AO-Grasp: Articulated Object Grasp Generation
Carlota Parés-Morlans (Stanford University), Jeannette Bohg (Stanford University)
GenerationPose EstimationRobotic IntelligencePoint Cloud
🎯 What it does: Propose the AO-Grasp method, which generates operable 6 DoF grasp points to interact with articulated objects
APEX: Ambidextrous Dual-Arm Robotic Manipulation Using Collision-Free Generative Diffusion Models
Apan Dastider, Mingjie Lin
Robotic IntelligenceDiffusion model
🎯 What it does: Propose the APEX method, utilizing a collision-free latent diffusion model to achieve motion planning and manipulation for dual-arm robots.
Applying Neural Monte Carlo Tree Search to Unsignalized Multi-intersection Scheduling for Autonomous Vehicles
Yucheng Shi, Vinny Cahill
Autonomous DrivingOptimizationReinforcement Learning
🎯 What it does: This paper applies Neural Monte Carlo Tree Search (NMCTS) to queue scheduling for autonomous vehicles at unsignalized multi-way intersections.
ARCADE: Scalable Demonstration Collection and Generation via Augmented Reality for Imitation Learning
Yue Yang, D. Szafir
Data SynthesisRobotic Intelligence
🎯 What it does: Propose the ARCADE framework, leveraging augmented reality to achieve robot demonstration collection and generation, enhancing the scalability of demonstration collection.
Archie Jnr: A Robotic Platform for Autonomous Cane Pruning of Grapevines
Henry Williams, Bruce A. MacDonald
Robotic IntelligenceImageAgriculture Related
🎯 What it does: Developed a robot platform named Archie Jnr capable of autonomously assessing grapevine structures and pruning low-quality branches like an expert pruner.
Archie Snr: A Robotic Platform for Autonomous Apple Fruitlet Thinning
Henry Williams, Bruce A. MacDonald
Robotic IntelligenceAgriculture Related
🎯 What it does: Developed a robot platform named Archie Snr capable of autonomously assessing the current fruit load of trees and removing excess apples as proficiently as professional thinners.
Architectural-Scale Artistic Brush Painting with a Hybrid Cable Robot
Gerry Chen, Seth Hutchinson
Robotic Intelligence
🎯 What it does: Built a hybrid cable-driven parallel robot combined with a 4-degree-of-freedom serial manipulator capable of painting 27m×3.7m murals on building facades;
ARDuP: Active Region Video Diffusion for Universal Policies
Shuaiyi Huang, Abhinav Shrivastava
GenerationRobotic IntelligenceVision-Language-Action ModelDiffusion modelOptical FlowVideoText
🎯 What it does: Proposed a video-based policy learning framework called ARDuP, emphasizing the generation of video plans that define goals through text to create active regions, thereby improving the generation of control actions.
Are Large Language Models Aligned with People’s Social Intuitions for Human–Robot Interactions?
Lennart Wachowiak, Gerard Canal
Robotic IntelligenceTransformerLarge Language ModelVision-Language-Action ModelVideoText
🎯 What it does: By replicating three human-robot interaction (HRI) user studies, comparing large language model (LLM) outputs with real participants' answers to evaluate whether LLMs can capture human social intuitions regarding judgments of robot behavior and communication preferences.
Arm-Constrained Curriculum Learning for Loco-Manipulation of a Wheel-Legged Robot
Zifan Wang, Guyue Zhou
Robotic IntelligenceReinforcement Learning
🎯 What it does: Integrate a robotic arm into a wheeled-legged robot and train control strategies using curriculum learning with arm constraints
AS-LIO: Spatial Overlap Guided Adaptive Sliding Window LiDAR-Inertial Odometry for Aggressive FOV Variation
Tianxiang Zhang, You Li
Autonomous DrivingRobotic IntelligenceSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposes an adaptive sliding window LiDAR-inertial odometry framework (AS-LIO) based on spatial overlap degree to improve localization accuracy and robustness in high-speed and drastic FOV change scenarios.
ASI-Seg: Audio-Driven Surgical Instrument Segmentation with Surgeon Intention Understanding
Zhen Chen, Hongbin Liu
SegmentationTransformerPrompt EngineeringVision-Language-Action ModelContrastive LearningImageMultimodalityBiomedical DataAudio
🎯 What it does: Developed ASI-Seg, an audio-driven surgical instrument segmentation framework;
ASML-VDIO: Visual-Depth-Inertial Odometry using Selected Accurate and Stable Multi-Modal Landmarks in Structural Environments
Xingjian Luo, Zheng Fang
Pose EstimationSimultaneous Localization and MappingMultimodality
🎯 What it does: Proposes the ASML-VDIO framework, which integrates RGB-D and IMU sensors, and improves the accuracy and efficiency of pose estimation through accurate and inaccurate classification of 3D landmarks and dynamic line feature removal methods.
Assessing Monocular Depth Estimation Networks for UAS Deployment in Rainforest Environments
Sri Sai Anirudh Tangellapalli, Brittany A. Duncan
SegmentationDepth EstimationConvolutional Neural NetworkImageVideoAgriculture Related
🎯 What it does: Evaluate monocular depth estimation models for assisting UAS data collection and navigation in rainforest environments, design a canopy height distinction segmentation pipeline based on MiDaS, and deploy example applications on edge systems.
ASY-VRNet: Waterway Panoptic Driving Perception Model based on Asymmetric Fair Fusion of Vision and 4D mmWave Radar
Runwei Guan, Yutao Yue
Object DetectionSegmentationAutonomous DrivingImageMultimodalityPoint CloudBenchmark
🎯 What it does: Propose an asymmetric fair fusion model ASY-VRNet based on vision and 4D millimeter-wave radar for waterway panoramic driving perception;
Asymptotically Optimal Lazy Lifelong Sampling-based Algorithm for Efficient Motion Planning in Dynamic Environments
Lu Huang, Xingjian Jing
Autonomous DrivingOptimization
🎯 What it does: Propose a progressively optimal lifelong sampling-based path planning algorithm that combines the advantages of lifelong planning and lazy search to quickly replan in dynamic environments, suitable for scenarios with high evaluation costs.
Asynchronous Event-Inertial Odometry using a Unified Gaussian Process Regression Framework
Xudong Li, Panfeng Huang
Pose EstimationOptimizationSimultaneous Localization and MappingMultimodality
🎯 What it does: Propose an asynchronous event-inertial odometry method based on a unified Gaussian process regression framework
Asynchronous Microphone Array Calibration using Hybrid TDOA Information
Chengjie Zhang, He Kong
OptimizationSimultaneous Localization and MappingAudio
🎯 What it does: Propose a batch SLAM method using hybrid TDOA (TDOA-S and TDOA-M) combined with kinematic information for asynchronous microphone array calibration.
Asynchronous Spatial-Temporal Allocation for Trajectory Planning of Heterogeneous Multi-Agent Systems
Yuda Chen, Zhongkui Li
Autonomous DrivingOptimization
🎯 What it does: Proposed an asynchronous spatiotemporal allocation method for trajectory planning in large-scale heterogeneous drone swarms, addressing the scalability limitations of synchronous distributed methods caused by the absence of global clock synchronization.
Attainable Force Approximation and Full-Pose Tracking Control of an Over-Actuated Thrust-Vectoring Modular Team UAV
Yen-Cheng Chu, Feng-Li Lian
Robotic Intelligence
🎯 What it does: Studied the achievable force space of modular UAV teams based on thrust vectorization under different configurations, and proposed an approximate feasible force space and a full attitude tracking controller.
Attitude Control of the Hydrobatic Intervention AUV Cuttlefish using Incremental Nonlinear Dynamic Inversion
Tom Slawik (DFKI RIC), F. Kirchner (DFKI RIC)
Autonomous Driving
🎯 What it does: Proposed an attitude control scheme for autonomous underwater vehicles (AUV) based on incremental nonlinear dynamic inversion (INDI) to achieve a 90-degree pitch ascent attitude transition.
Audio-Visual Traffic Light State Detection for Urban Robots
Sagar Gupta, Akansel Cosgun
Object DetectionRobotic IntelligenceImageMultimodalityAudio
🎯 What it does: Proposes a multimodal traffic light state detection method based on vision and audio, applied to quadruped robots in urban environments.
Augmenting Vision with Radar for All-weather Geo-localization without a Prior HD Map
Can Dong, Huijun Gao
Autonomous DrivingSimultaneous Localization and MappingImageMultimodality
🎯 What it does: Proposes the first method based on camera and radar fusion that achieves robust geolocation under all weather conditions.
AutoInst: Automatic Instance-Based Segmentation of LiDAR 3D Scans
Cedric Perauer, A. Artemov
SegmentationAutonomous DrivingContrastive LearningPoint Cloud
🎯 What it does: Propose an unsupervised LiDAR 3D scanning instance segmentation method that first generates initial instance masks through pseudo annotations and then refines the segmentation results using a self-training algorithm.
AutoJoin: Efficient Adversarial Training against Gradient-Free Perturbations for Robust Maneuvering via Denoising Autoencoder and Joint Learning
Michael Villarreal, Weizi Li
Adversarial AttackAuto EncoderImage
🎯 What it does: Proposed a gradient-free adversarial training method called AutoJoin for image-driven manipulation tasks.
Automatic 3D Road Surface Reconstruction via Cross-Section Modeling and Interpolation*
Matteo Bellusci, Matteo Matteucci
Autonomous DrivingPoint Cloud
🎯 What it does: Using data collected by vehicle-mounted multi-sensors (including LiDAR), first obtain the clothoidal representation of the road surface boundary, then extract and interpolate smooth 3D cross-sectional curves to generate an analytical-form 3D road surface that can be rendered in detail at any resolution.
Automatic design of robot swarms that perform composite missions: an approach based on inverse reinforcement learning
Jeanne Szpirer, Mauro Birattari
Robotic IntelligenceReinforcement Learning from Human FeedbackReinforcement Learning
🎯 What it does: This paper extends inverse reinforcement learning (IRL) to multi-objective optimization, automatically designing control software for robot swarms to execute composite tasks, and simulates 12 composite tasks on 20 e-puck robots.