arXivSub Start free trial

IROS 2024 Papers with AI Summaries

IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers

(Real2Sim)−1: 3D Branch Point Cloud Completion for Robotic Pruning in Apple Orchards

Tian Qiu, Yu Jiang

Data SynthesisDomain AdaptationRobotic IntelligencePoint CloudAgriculture Related

🎯 What it does: Develop a closed-loop (Real2Sim)-1 method that generates realistic 3D apple tree models through a Real-to-Simulation data generation pipeline without manual parameterization, and uses this data to train a deep model for point cloud completion and skeletonization of real partial branch point clouds, thereby improving robotic pruning accuracy.

2mm Diameter Continuum Robot Tools for Suturing in Open Spina Bifida Repair

A. Law, T. Looi

Robotic Intelligence

🎯 What it does: Designed a 3-degree-of-freedom, 2mm-diameter continuous robot tool wrist for suturing in open spina bifida repair.

3D Affordance Keypoint Detection for Robotic Manipulation

Zhiyang Liu, F. E. Tay

Robotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Proposes FAKP-Net, a 3D keypoint network that integrates RGB and depth images to detect manipulable keypoints and their corresponding execution positions, orientations, and ranges, supporting robotic manipulation;

3D Global Path Planning for Walking Robots on Sparse Volumetric Maps

M. G. Besselmann, R. Dillmann

OptimizationRobotic Intelligence

🎯 What it does: Proposes a method for global path planning of walking robots on complete 3D volume maps, achieving optimal path search in 3D space through an extension of the traditional A* algorithm.

3D Localization of Objects Buried within Granular Material Using a Distributed 3-Axis Tactile Sensor

Zheng Chen, Lorenzo Jamone

Pose EstimationRecurrent Neural NetworkTime Series

🎯 What it does: Propose a method using distributed 3-axis tactile sensors (which simultaneously measure normal and shear forces) and employ three LSTM models for real-time 3D localization of objects buried in granular media, determining their distance, direction, and depth.

3D Object Detection via Stereo Pyramid Transformers with Rich Semantic Feature Fusion

Rongqi Gu, Guang Chen

Object DetectionAutonomous DrivingTransformerImage

🎯 What it does: Proposed a Transformer-based stereo pyramid Transformer backbone network, which integrates stereo geometric volumes (SPSV, MSV, LV) to enhance feature representation, thereby achieving 3D object detection in stereo images.

3D Object Visibility Prediction in Autonomous Driving

Chuanyu Luo, Pu Li

Autonomous Driving

🎯 What it does: Proposed the 3D object visibility attribute and its calculation algorithm, integrating them into a perception model as part of a multi-task learning framework.

3D Ultrasound Image Acquisition and Diagnostic Analysis of the Common Carotid Artery with a Portable Robotic Device

Longyue Tan, Shuangyin Wang

SegmentationRobotic IntelligenceBiomedical DataUltrasound

🎯 What it does: A semi-automated method for carotid ultrasound image acquisition based on a portable robotic device is proposed, utilizing collected posture, force sensing, and synchronized ultrasound data for 3D reconstruction and intima-media thickness (IMT) segmentation.

3D-BLUE: Backscatter Localization for Underwater Robotics

Sayed Saad Afzal, Fadel M. Adib

Robotic IntelligenceSimultaneous Localization and MappingAudio

🎯 What it does: Designed, implemented, and evaluated 3D-BLUE, an ultra-low power 3D positioning system suitable for compact underwater robots that can achieve accurate positioning in shallow water environments.

3DGS-Calib: 3D Gaussian Splatting for Multimodal SpatioTemporal Calibration

Quentin Herau, C. Demonceaux

Autonomous DrivingGaussian SplattingMultimodality

🎯 What it does: Proposes 3DGS-Calib, a method that uses 3D Gaussian Splatting to achieve multi-modal spatiotemporal calibration.

3DR-DIFF: Blind Diffusion Inpainting for 3D Point Cloud Reconstruction and Segmentation

K.T.Yasas Mahima, Matt Garratt

RestorationSegmentationDiffusion modelPoint Cloud

🎯 What it does: Propose a blind completion method based on diffusion networks, 3DR-DIFF, for reconstructing and segmenting damaged LiDAR point clouds.

6-DoF Grasp Detection in Clutter with Enhanced Receptive Field and Graspable Balance Sampling

Hanwen Wang, Jian Li

Pose EstimationRobotic IntelligencePoint Cloud

🎯 What it does: This paper proposes a 6-DoF grasp detection method with enhanced receptive field for small-scale grasping, and designs a grasp balance sampling module based on a 3D segmentation network to improve the recognition capability of small object grasping.

6D Variable Virtual Fixtures for Telemanipulated Insertion Tasks

Stephan Andreas Schwarz, Ulrike Thomas

Robotic IntelligenceImageBiomedical Data

🎯 What it does: Proposed a 6D variable virtual device for remote manipulation of insertion tasks, employing virtual guidance, position constraints, distance-related variable stiffness, and energy tank stability assurance; using nasopharyngeal sampling as a case study;

A ’MAP’ to find high-performing soft robot designs: Traversing complex design spaces using MAP-elites and Topology Optimization

Yue Xie, Fumiya Iida

OptimizationRobotic Intelligence

🎯 What it does: Developed a hierarchical design optimization method that combines topology optimization with MAP-elites (quality diversity optimization) to evolve variable-sized void regions in soft robot design.

A 6-DOF Double-layer Programmable Remote Center of Motion Robot for Vitreoretinal Surgery

Chenyu Wang, S. Ko

OptimizationRobotic Intelligence

🎯 What it does: Designed and implemented a 6-degree-of-freedom (DOF) retinal surgery robot with a dual-layer programmable remote center of motion (RCM) mechanism, utilizing dual-plane 5-bar linkage combinations, genetic algorithm-optimized workspace, and experimental validation on a transparent eyeball model.

A Biomimetic Robot Crawling Upstream using Adhesive Suckers Inspired by Net-winged Midge Larvae

Haoyuan Xu, Li Wen

Robotic Intelligence

🎯 What it does: Designed and tested a bio-inspired suction cup with ridge-shaped micro-hair structures, integrating it into a quadrupedal crawling robot capable of walking in both air and water, achieving upstream movement in fast-flowing water;

A Cascaded Broad Learning System for Manipulator Motion Control

Guoyu Zuo, Shuangyue Yu

Robotic Intelligence

🎯 What it does: Proposed a robotic arm motion control strategy based on the Cascading Feature Enhanced Elastic Net Generalized Learning System (CFE-EN-BLS)

A Case Study on Visual-Audio-Tactile Cross-Modal Retrieval

Jagoda Wojcik, Shan Luo

RetrievalMultimodality

🎯 What it does: Propose a cross-modal retrieval model VAT-CMR that integrates visual, audio, and tactile multimodal information;

A Circular Soft Pneumatic Actuator with Bi-Directional Bending Behavior

Jeannette Circe, M. Rosen

Robotic Intelligence

🎯 What it does: Designed, manufactured, and analyzed a circular soft actuator that achieves bidirectional bending using only single-chamber positive pressure.

A Closed-loop Control for Lower Limb Exoskeleton Considering Overall Deformations: A Simple and Direct Application Method

Feng Li, Xinyu Wu

Robotic IntelligenceTime Series

🎯 What it does: Proposes a closed-loop control method combining the Deformation Relation Network (DRN) with a Fractional-Order Viscoelastic (FOV) controller to compensate for full deformation and achieve stable dynamic walking.

A Collaborative Stereo Camera with Two UAVs for Long-distance Mapping of Urban Buildings

Zhaoying Wang, Wei Dong

Simultaneous Localization and MappingImage

🎯 What it does: Using two UAVs equipped with cameras to construct a wide-baseline collaborative stereo vision system, enabling the creation of navigable mesh maps for distant urban buildings;

A comparison of audible, visual, and multi-modal communication for multi-robot supervision and situational awareness

Richard Attfield, Dana Kulić

Robotic IntelligenceImageMultimodalityAudio

🎯 What it does: Compared the effects of unimodal and multimodal audio-visual nonverbal cues on situation awareness in multi-robot supervision

A Comprehensive Modeling and Scheduling Approach for Allocating Distributed Multi-Robot Software to the Edge/Cloud

Yongzhou Zhang, Björn Hein

OptimizationRobotic Intelligence

🎯 What it does: This paper proposes a comprehensive modeling and scheduling method for distributed multi-robot systems, addressing the selection and allocation of software modules. By constructing a taxonomy of software modules and a computing cluster model, the scheduling problem is formulated as a multi-objective optimization problem.

A Contact Model based on Denoising Diffusion to Learn Variable Impedance Control for Contact-rich Manipulation

Masashi Okada, Tadahiro Taniguchi

Robotic IntelligenceDiffusion model

🎯 What it does: Proposed and implemented a contact model based on denoising diffusion models (DCM) to learn variable damping control for robots in contact-rich tasks such as wiping, reducing experimental costs by predicting robotic contact trajectories.

A Context-Enhanced Full-Resolution Floor Plan Segmentation Network for Topological Semantic Mapping

Zhengcai Cao, Meng Zhou

SegmentationConvolutional Neural NetworkImage

🎯 What it does: Proposed a Context-Enhanced Full-Resolution Network (CEFRN) for semantic segmentation of indoor floor plans, and applied it to topological semantic maps for partially visually impaired individuals.

A Convex Formulation of Frictional Contact for the Material Point Method and Rigid Bodies

Zeshun Zong, Xuchen Han

OptimizationPhysics Related

🎯 What it does: Proposed a convex optimization framework that unifies the Material Point Method (MPM) with rigid body kinematics under frictional contact scenarios

A Cooperative Recovery Framework for Resilient Multi-Robot Swarm Operations Under Loss of Localization in Unknown Environments

Paul J. Bonczek, N. Bezzo

Anomaly DetectionRobotic Intelligence

🎯 What it does: A decentralized framework is proposed, enabling robots to detect anomalies when losing localization capabilities and notify neighbors through information exchange. Neighbors utilize their own mobility to assist affected robots in relocalization.

A CT-guided Control Framework of a Robotic Flexible Endoscope for the Diagnosis of the Maxillary Sinusitis

P. Zhu, Kwok Wai Samuel Au

Robotic IntelligenceBiomedical DataComputed Tomography

🎯 What it does: Proposed a CT-guided robotic flexible endoscope control framework for diagnosing sinusitis; designed a feasible path to the target location and proposed an optimal control scheme, enabling the robot to autonomously follow the path and reach the target location; conducted experiments on a widely used sinus phantom model, verifying the framework's accuracy and feasibility.

A Data-Informed Analysis of Scalable Supervision for Safety in Autonomous Vehicle Fleets

C. Hickert, Cathy Wu

Autonomous DrivingTabularTime Series

🎯 What it does: Propose the DISCES framework, which models the vehicle arrival process at 1097 highway merge points using microscopic traffic reconstruction and queueing theory, analyzes the scalability of remote supervision in safety-critical scenarios, and provides two methods to reduce supervision demand.

A Decentralized Partially Observable Markov Decision Process for Dynamic Obstacle Avoidance and Complete Area Coverage using Multiple Reconfigurable Robots

J. Pey, M. R. Elara

Reinforcement Learning

🎯 What it does: Proposed a multi-agent reinforcement learning framework based on Dec-POMDP, enabling reconfigurable robots to achieve complete area coverage in dynamic obstacle environments.

A Deep Reinforcement Learning Framework and Methodology for Reducing the Sim-to-Real Gap in ASV Navigation

L. F. W. Batista, Cédric Pradalier

Domain AdaptationAutonomous DrivingReinforcement Learning

🎯 What it does: Integrate buoyancy and fluid dynamics models into a modern reinforcement learning framework, and enhance the performance of RL agents through system identification and domain randomization, with real-world application for collecting floating debris tasks.

A Deep Signed Directional Distance Function for Shape Representation

Ehsan Zobeidi, Nikolay Atanasov

GenerationRepresentation Learning

🎯 What it does: Proposes the Signed Direction Distance Function (SDDF) and the DeepSDDF deep neural network based on it, for capturing view information and achieving fast, accurate synthesis from new perspectives;

A Digital Twin-Driven Immersive Teleoperation Framework for Robot-Assisted Microsurgery

Peiyang Jiang, Dandan Zhang

Robotic IntelligenceBiomedical Data

🎯 What it does: Established an immersive teleoperation framework based on digital twin and mixed reality for robot-assisted minimally invasive surgery

A Direct Algorithm for Multi-Gyroscope Infield Calibration

Tianheng Wang, S. Roumeliotis

OptimizationComputational EfficiencyPhysics Related

🎯 What it does: This paper proposes a direct algorithm based on least squares for estimating the rotation extrinsic parameters and scale factors of two rigidly mounted gyroscopes on the same device, without requiring any iteration;

A Direct Semi-Exhaustive Search Method for Robust, Partial-to-Full Point Cloud Registration

Richard Cheng, Mark Tjersland

Pose EstimationPoint Cloud

🎯 What it does: Proposes the Direct Semi-Exhaustive Search (DSES) method for robust point cloud registration without corresponding points

A Ducted Fan UAV for Safe Aerial Grabbing and Transfer of Multiple Loads Using Electromagnets

Zhong Yin, Hailong Pei

Robotic Intelligence

🎯 What it does: Developed a coaxial tube fan drone equipped with an electromagnet to achieve safe aerial grasping and transferring of multiple payloads, supporting direct human-robot cargo transfer

A Facile one-step injection novel composite sensor for robot tactile assistance

Yuyin Zhang, Toshio Fukuda

Robotic Intelligence

🎯 What it does: A thin film composite sensor based on liquid metal and multi-layer PDMS microfluidic layers was fabricated using a one-time injection process, enabling simultaneous and decoupled measurement of pressure and deformation to provide tactile assistance for robots in surgical procedures and manipulation of human tissues.

A Fast Motion and Foothold Planning Framework for Legged Robots on Discrete Terrain

Jiyu Yu, Rong Xiong

OptimizationRobotic Intelligence

🎯 What it does: Proposed a real-time motion and foot placement planning framework for discrete terrains, which includes global path planning, local foot position optimization, and an execution layer based on MPC and WBC;

A Fast Online Omnidirectional Quadrupedal Jumping Framework Via Virtual-Model Control and Minimum Jerk Trajectory Generation

Linzhu Yue, Yunhui Liu

OptimizationRobotic Intelligence

🎯 What it does: Proposed a fast omnidirectional jumping framework enabling quadruped robots to achieve rapid, real-time omnidirectional jumps.

A Feasibility Study of a Soft, Low-Cost, 6-Axis Load Cell for Haptics

Madison Veliky, Nabil Simaan

Robotic Intelligence

🎯 What it does: Designed and calibrated a low-cost, flexible 6-axis force/torque sensor specifically for laparoscopic haptic training applications.

A Framework for Neurosymbolic Goal-Conditioned Continual Learning in Open World Environments

Pierrick Lorang, Matthias Scheutz

Robotic IntelligenceReinforcement Learning

🎯 What it does: Proposed an integrated Task and Motion Planning (TAMP) architecture combining symbolic planning with reinforcement learning to achieve unsupervised continuous learning and adaptation.

A Framework for Reproducible Benchmarking and Performance Diagnosis of SLAM Systems

Nikola Radulov, Mikel Luján

Simultaneous Localization and MappingMultimodalityBenchmark

🎯 What it does: Proposes SLAMFuse, an open-source benchmark framework for multi-modal SLAM algorithms, providing a consistent cross-platform environment, data fuzz testing, failure detection, and diagnostic tools;

A General Formulation for Path Constrained Time-Optimized Trajectory Planning with Environmental and Object Contacts

Dasharadhan Mahalingam, Nilanjan Chakraborty

OptimizationRobotic Intelligence

🎯 What it does: Proposes a path-constrained time-optimal trajectory planning problem considering nonlinear friction cone constraints from hand-object and object-environment contact, and provides a solution method based on second-order cone programming (SOCP).

A Generic Trajectory Planning Method for Constrained All-Wheel-Steering Robots

Ren Xin, Ming Liu

OptimizationRobotic Intelligence

🎯 What it does: Proposes a general trajectory planning method for four-wheel steering (AWS) robots with fixed steering axes and limited steering angles for each wheel, called C-AWS.

A Geometry-based Approach for Support-free Additive Manufacturing of Structures with Large Overhang Angles and Closed Features

Jitian Liu, Michael D. M. Kutzer

🎯 What it does: Propose a free additive manufacturing method that can fabricate 3D structures with large overhang angles and fully enclosed cavities without using support structures.

A Graph-Based Self-Calibration Technique for Cable-Driven Robots with Sagging Cable

M. R. Dindarloo, H. Taghirad

OptimizationRobotic Intelligence

🎯 What it does: Proposed a graph-based self-calibration framework specifically addressing cable sagging issues in large cable-driven parallel robots, relying solely on internal sensors for calibration;

A Heterogeneous System of Systems Framework for Proactive Path Planning of a UAV-assisted UGV in Uncertain Environments

Patrick Sherman, N. Bezzo

Autonomous DrivingOptimizationRobotic Intelligence

🎯 What it does: Propose a UAV-assisted heterogeneous system architecture, utilizing UAVs for front reconnaissance and predicting UGV's reachable states through sampling virtual particles, thereby proactively evaluating and updating UGV path planning, and using MPPI to optimize paths.

A High-Performance Anthropomorphic Robotic Arm for Household Applications

Tianliang Liu, Yu Zheng

Robotic Intelligence

🎯 What it does: Propose a high-performance humanoid robotic arm TRX-Arm based on differential cable transmission, and experimentally verify its high repeatability, high payload capacity, and high dynamic performance on a dual-arm platform.

A Hybrid Human Tracking System using UWB Sensors and Monocular Visual Data Fusion for Human Following Robots

Dingzhi Zhang, Tim Lüth

Object TrackingRobotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Propose a hybrid human tracking system combining UWB sensors and monocular visual localization for human-following robots.

A Hybrid Model and Learning-Based Force Estimation Framework for Surgical Robots

Hao Yang, Jie Ying Wu

Robotic Intelligence

🎯 What it does: Proposed a hybrid framework combining model-based and learning methods for surgical robotic force estimation, specifically targeting the Patient Side Manipulators of the da Vinci research platform.

A Hybrid Vision/Force Control Strategy for Handheld Robotic Devices Enhancing Probe-Based Confocal Laser Endomicroscopy

Ingu Choi, Sungwook Yang

Robotic IntelligenceBiomedical Data

🎯 What it does: Introduces a hybrid visual/force control strategy for handheld robotic devices to obtain clear and consistent images during probe-based confocal laser endomicroscopy imaging.

A Language-Driven Navigation Strategy Integrating Semantic Maps and Large Language Models

Zhengjun Zhong, Fei Ma

Autonomous DrivingLarge Language Model

🎯 What it does: Developed a language-driven navigation strategy that integrates foundation model semantic segmentation, semantic maps, and large language models

A Large Vision-Language Model based Environment Perception System for Visually Impaired People

Zezhou Chen, Shiguo Lian

Object DetectionSegmentationTransformerLarge Language ModelVision Language ModelImageMultimodality

🎯 What it does: Developed an environment perception system based on large vision-language models (LVLM) to help visually impaired individuals obtain global descriptions of their surroundings, object classification, and detailed information about target objects through wearable devices.

A Laser-Induced Graphene-Based Flexible Multimodal Sensor for Material and Texture Perception

Youning Duo, Li Wen

ClassificationRecognitionConvolutional Neural Network

🎯 What it does: Developed a flexible multimodal sensor based on laser-induced graphene, capable of simultaneously detecting non-contact signals (distance, material) and tactile signals (3D force), and utilizing CNN to achieve recognition of different materials and fabric textures.

A Learning-based Controller for Multi-Contact Grasps on Unknown Objects with a Dexterous Hand

Dominik Winkelbauer, Berthold Bäuml

Robotic Intelligence

🎯 What it does: Proposed a learning-based grasping controller that supports multi-contact grasping on unknown objects.

A Lightweight De-confounding Transformer for Image Captioning in Wearable Assistive Navigation Device

Zhengcai Cao, Meng Zhou

GenerationComputational EfficiencyTransformerVision Language ModelImageText

🎯 What it does: Proposed a lightweight deconfounding Transformer network (LDTNet) for image caption generation in wearable assistive navigation devices;

A Low-Cost, High-Speed, and Robust Bin Picking System for Factory Automation Enabled by a Non-stop, Multi-View, and Active Vision Scheme

Xingdou Fu, M. Suwa

Robotic IntelligencePoint Cloud

🎯 What it does: Designed a low-cost, high-speed, and robust box-picking system using a non-stop multi-view active vision approach, placing sensors on the arm to achieve tight coupling with robotic motion tasks.

A Low-Texture Robust Hybrid Feature Based Visual Odometry

He Wang, Ru Li

Simultaneous Localization and MappingOptical FlowImage

🎯 What it does: Developed a visual odometry system for low-texture scenarios, integrating vanishing point optimization for camera pose and Manhattan-assisted optical flow for matching line segments.

A mathematical characterization of the convergence domain for Direct Visual Servoing

Meriem Belinda Naamani, E. Mouaddib

Robotic Intelligence

🎯 What it does: Proposed a mathematical representation of the convergence domain of DVS and provided a closed-form expression for the controlled degrees of freedom.

A Mixed-Integer Conic Program for the Moving-Target Traveling Salesman Problem based on a Graph of Convex Sets

Allen George Philip, H. Choset

OptimizationComputational Efficiency

🎯 What it does: Propose a new mixed-integer cone programming (MICP) formulation that leverages the convex set formed by the trajectory of a moving target in a spatiotemporal coordinate system, transforming the problem of finding the optimal solution for the mobile target traveling salesman problem (MT-TSP) into identifying the shortest path in a convex set graph that satisfies velocity constraints.

A Multi-model Fusion of LiDAR-inertial Odometry via Localization and Mapping

An Duy Nguyen, H. M. La

Autonomous DrivingSimultaneous Localization and MappingPoint Cloud

🎯 What it does: Proposed a LiDAR-IMU fusion odometry framework that utilizes IMMKF to achieve reliable navigation in dynamic motion and noisy environments, and employs ikd-tree to realize continuous map building.

A Neurosymbolic Approach to Adaptive Feature Extraction in SLAM

Yasra Chandio, F. Anwar

Simultaneous Localization and Mapping

🎯 What it does: Propose a neuro-symbolic method for adaptive feature extraction in the feature extraction module of SLAM.

A New 10-mg SMA-Based Fast Bimorph Actuator for Microrobotics

Conor K. Trygstad, N. O. Pérez-Arancibia

Robotic Intelligence

🎯 What it does: Developed a 10 mg, 14 mm long, 4.8 mm³ volume SMA-based bipolar configuration rapid actuator for microrobots.

A Non-Homogeneity Mapless Navigation Based on Hierarchical Safe Reinforcement Learning in Dynamic Complex Environments

Jianmin Qin, Jiahu Qin

Autonomous DrivingReinforcement Learning

🎯 What it does: Proposed a mapless navigation method called HSRLN based on hierarchical safe reinforcement learning, which trains a mapless navigation strategy applicable to non-uniform complex scenarios.

A non-invasive device for skin cancer diagnosis: first clinical evidence with spectroscopic data enhanced by machine learning algorithms

†. V.Mainardi, G. Ciuti

ClassificationBiomedical Data

🎯 What it does: Developed a non-invasive device that uses near-infrared reflectance spectroscopy technology to analyze skin lesions and employs machine learning algorithms to classify different types of skin lesions;

A Novel Approach for Precise Tissue Tracking in Breast Lumpectomy

Yeganeh Aliyari, Mahdi Tavakoli

Graph Neural NetworkAuto EncoderBiomedical Data

🎯 What it does: A new method for precise tissue tracking during breast surgery was studied, utilizing a real tissue deformation simulator and AI technology to accurately locate tumors, thereby enhancing surgical precision and safety.

A Novel Framework for Structure Descriptors-Guided Hand-drawn Floor Plan Reconstruction

Zhentong Zhang, Shaokun Zhang

RestorationRepresentation LearningRobotic IntelligenceImage

🎯 What it does: Propose a new framework for identifying abstract hand-drawn elements and achieving standardized reconstruction of hand-drawn floor plans, providing global map information to robots

A Novel Geometrical Structure Robot Hand for Linear-parallel Pinching and Coupled Self-adaptive Hybrid Grasping

Shi Chen, Wenzeng Zhang

Robotic Intelligence

🎯 What it does: Designed and analyzed an LPCSA robotic hand with three modes: linear parallel pinch grip, coupled grasp, and adaptive grasp.

A Novel Variable Step-size Path Planning Framework with Step-Consistent Markov Decision Process For Large-Scale UAV Swarm

Dan Xu, Chang Wang

Autonomous DrivingOptimizationReinforcement Learning

🎯 What it does: This paper proposes a novel training framework named RVSDRL for large-scale drone formations and variable-step path planning, combining centralized training with ground servers and decentralized distributed training for drones;

A Novel Variable Stiffness Suspension System for Improved Stability and Control of Tactile Mobile Manipulators

Sebastian Kuhn, Sami Haddadin

Robotic Intelligence

🎯 What it does: Designed and verified a variable stiffness suspension system for a mobile base to enhance the stability and safety of collaborative mobile manipulators in dynamic environments.

A Novel Vitreoretinal Surgical Robot System to Maximize the Internal Reachable Workspace and Minimize the External Link Motion*

Gowoon Jeong, Seong Yeong Ko

Robotic Intelligence

🎯 What it does: A novel ophthalmic minimally invasive surgical robot system was developed, integrating coaxial tubes (CTM) and belt-driven remote center of motion (RCM) mechanisms to maximize the internal reachable workspace and minimize external mechanical movement.

A parallel-actuated robot with two end-effector degrees-of-freedom: Application as a novel wearable head-neck traction brace

Jingzong Zhou, Sunil K. Agrawal

Robotic Intelligence

🎯 What it does: A parallel-driven robot mechanism was designed to provide two degrees of freedom to the end effector, applied to a cervical traction frame.

A Perceptive Pneumatic Artificial Muscle Empowered by Double Helix Fiber Reinforcement

Yufeng Wang, Hongbo Wang

Robotic Intelligence

🎯 What it does: Developed a dual-helix fiber-reinforced pneumatic artificial muscle (PPAM), which achieves elongation by restricting the radial expansion of the cylindrical body and enables real-time length sensing through inductance measurement; established a finite element model and derived theoretical formulas; conducted rigorous experimental validation.

A Piecewise-weighted RANSAC Method Utilizing Abandoned Hypothesis Model Information with a New Application on Robot Self-calibration

Jianhui He, Junjie Li

Robotic Intelligence

🎯 What it does: A robot self-calibration method based on segmented weighted RANSAC is proposed to reduce non-kinematic errors introduced by two types of pose samples during self-calibration.

A Point-Based Approach to Efficient LiDAR Multi-Task Perception

Christopher Lang, Abhinav Valada

Object DetectionSegmentationAutonomous DrivingComputational EfficiencyTransformerPoint Cloud

🎯 What it does: Propose an efficient point-based multi-task perception network called PAttFormer for point cloud semantic segmentation and object detection;

A Point-Line Features Fusion Method for Fast and Robust Monocular Visual-Inertial Initialization

Guoqiang Xie, Yiguang Liu

Pose EstimationSimultaneous Localization and MappingImageVideo

🎯 What it does: Proposed a point-line feature fusion method for monocular visual-inertial initialization

A Proxy-Tactile Reactive Control for Robots Moving in Clutter

Giammarco Caroleo, P. Maiolino

Robotic Intelligence

🎯 What it does: Proposed a reactive control method that simultaneously utilizes proximity sensing and tactile feedback to execute grasp-and-place tasks in unknown cluttered environments.

A Real-time Filter for Human Pose Estimation based on Denoising Diffusion Models for Edge Devices

Chiara Bozzini, N. Bombieri

Pose EstimationComputational EfficiencyConvolutional Neural NetworkDiffusion model

🎯 What it does: Propose a real-time filter based on the denoising diffusion model (DM) to enhance the results of CNN-based human pose estimation (HPE) on edge devices, achieving extremely low computational overhead.

A Robot Kinematics Model Estimation Using Inertial Sensors for On-Site Building Robotics

Hiroya Sato, Masayuki Inaba

Robotic Intelligence

🎯 What it does: Estimate the robot's kinematic model using an inertial measurement unit (IMU) on a rigid link.

A Robotic Mediation Device for Skill Assessment and Training During Colonoscopy

O. Richards, Nabil Simaan

ClassificationRobotic IntelligenceTime SeriesBiomedical Data

🎯 What it does: Developed a handheld robotic mediation device for skill assessment and training in colonoscopy operations, achieving objective evaluation by measuring the user's kinematic trajectories and power input;

A Robotic-centric Paradigm for 3D Human Tracking Under Complex Environments Using Multi-modal Adaptation

Shuo Xin, Yong Liu

Object TrackingRobotic IntelligenceTransformerVideoPoint Cloud

🎯 What it does: Proposed a robot-centric 3D human tracking paradigm called MATNet, which can directly match LiDAR point clouds with RGB videos to achieve end-to-end learning.

A Robust and Efficient Robotic Packing Pipeline with Dissipativity- Based Adaptive Impedance-Force Control

Zhenning Zhou, M. H. Ang

Robotic Intelligence

🎯 What it does: Proposes a robust and efficient robotic packaging pipeline, comprising external force estimation based on a generalized momentum observer, an adaptive variable damping dissipative controller, and a three-step packaging method tailored for different interaction states.

A Robust Visual SLAM System for Small-Scale Quadruped Robots in Dynamic Environments

Chengyang Li, Qing Shi

SegmentationRobotic IntelligenceSimultaneous Localization and Mapping

🎯 What it does: Designed a robust visual SLAM system called ViQu-SLAM for small quadruped robots in dynamic environments, aiming for precise localization and reducing data association errors caused by moving objects.

A Safe and Efficient Timed-Elastic-Band Planner for Unstructured Environments

Haoyue Xi, Yu Hu

Autonomous DrivingOptimizationSimultaneous Localization and Mapping

🎯 What it does: Proposed the Timed-Elastic-Band planner TEB-U for unstructured environments, incorporating a free space extraction optimization module, a dynamic global point adjustment module, and replacing traditional obstacle constraints with boundary constraints in hypergraph optimization to enhance safety and real-time performance.

A Sampling Ensemble for Asymptotically Complete Motion Planning with Volume-Reducing Workspace Constraints

Sihui Li, Neil T. Dantam

Robotic IntelligenceStochastic Differential EquationOrdinary Differential Equation

🎯 What it does: Developed a sampling set for workspace constraints, integrating SDCL, dual quaternion optimization projection, and random sampling to achieve efficient and robust motion planning.

A Scalable Platform for Robot Learning and Physical Skill Data Collection

Samuel Schneider, Sami Haddadin

OptimizationRobotic Intelligence

🎯 What it does: Developed an scalable robot learning and physical skill data collection platform PD.RAI, integrating robot learning units (RLUs), a global database, and a robot swarm intelligence system (RoCI), achieving parallelized, distributed tactile manipulation learning and data acquisition

A Series Variable-stiffness Joint for Robot-assisted Resistance Training

Xingyu Hu, Yanggang Feng

Robotic Intelligence

🎯 What it does: Proposed and prototyped a compact passive variable stiffness joint for robot-assisted resistance training, featuring adjustable stiffness and damping as well as passive angle tracking functionality.

A Service Robot in the Wild: Analysis of Users Intentions, Robot Behaviors, and Their Impact on the Interaction

Simone Arreghini, Antonio Paolillo

Robotic Intelligence

🎯 What it does: Studied a service robot that provides chocolate gifts to people, capable of predicting human interaction intentions and performing a 'delivery' gesture, while collecting real interaction data in various public scenarios;

A Slices Perspective for Incremental Nonparametric Inference in High Dimensional State Spaces

Moshe Shienman, V. Indelman

Computational Efficiency

🎯 What it does: Proposes a slicing perspective method for incremental nonparametric probabilistic inference in high-dimensional state spaces.

A Soft Robotic Finger Inspired by Biological Perception Models for Tactile Sensing

Baijin Mao, Juntian Qu

Robotic Intelligence

🎯 What it does: This paper proposes a bio-inspired soft mechanical finger with shape adaptability and multimodal tactile sensing capabilities. It employs an advanced Fin Ray structure and equips the fingertip with tactile bristles to perceive obstacle direction, fingertip pressure, surface roughness, and the size of grasped spheres. A distributed fiber optic sensing technology is used to construct a multi-point, multimodal tactile perception neural network. The fiber optic sensors are seamlessly integrated into the soft mechanical finger using 3D printing and silicone coating techniques, achieving intelligent perception. Experimental verification demonstrates that the soft mechanical finger exhibits high sensitivity and accuracy in force perception and curvature identification, with an accuracy rate reaching 100%.

A Soft Robotic System Automatically Learns Precise Agile Motions Without Model Information

Simon Bachhuber, T. Seel

Robotic IntelligenceTime SeriesOrdinary Differential Equation

🎯 What it does: Use ANODEC to automatically learn soft robots to perform agile non-repetitive reference tracking actions within a limited time.

A Tactile Lightweight Exoskeleton for Teleoperation: Design and Control Performance

Moein Forouhar, Sami Haddadin

Robotic Intelligence

🎯 What it does: Developed a lightweight haptic exoskeleton, improved the 3-degree-of-freedom (DOF) actuation of the shoulder joint and the Baldwin cable pre-tension adjustment mechanism, and installed force/torque sensors at the contact points to directly measure the interaction force between the human arm and the exoskeleton.

A Tetherless Soft Robotic Wearable Haptic Human Machine Interface for Robot Teleoperation

Shilpa Thakur, C. Onal

Robotic Intelligence

🎯 What it does: Developed and evaluated a cable-free soft wearable tactile human-machine interface for robotic teleoperation.

A Two-Stage Reinforcement Learning Approach for Robot Navigation in Long-range Indoor Dense Crowd Environments

Xinghui Jing, Long Zeng

Robotic IntelligenceReinforcement Learning

🎯 What it does: Propose a two-stage deep reinforcement learning method called TSDRL for mobile robot navigation in long-distance indoor crowded environments.

A Unified Framework of Hybrid Vision-Force Control With Nullspace Compliance for Redundant Robots

Zhiwen Li, Yongping Pan

Robotic Intelligence

🎯 What it does: Propose a unified hybrid visual-force control framework that integrates Cartesian space HVFC with joint clearance damping control into closed-loop dynamics, applicable to redundant robots without requiring contact surface geometry knowledge.

A Unified Interaction Control Framework for Safe Robotic Ultrasound Scanning with Human-Intention-Aware Compliance

Xiangjie Yan, Xiang Li

Robotic IntelligenceBiomedical DataUltrasound

🎯 What it does: Proposed a unified interactive control framework for safe robotic ultrasound scanning that can handle all common interaction scenarios.

A velocity dependent delayed output feedback control (v-DOFC) for gait assistance with an ergonomically designed bi-directional cable-driven hip assist device

Dong Hyun Kim, Minhyung Lee

Robotic Intelligence

🎯 What it does: Developed a 1.5 kg soft wearable bi-directional rope-driven hip exosuit with a velocity-related delayed output feedback controller to achieve flexible flexion and extension assistance.

A Voxel-Enabled Robotic Assistant for Omnidirectional Conveyance

M. Carvajal, T. Padır

Robotic Intelligence

🎯 What it does: Propose a Voxel-Enabled Robotic Assistant (VERA), achieving omnidirectional transport through networked, dynamically configurable workbench partitions and hot-swappable 'voxel' modules;

A Wearable Mechanical Pressure-Electrophysiological Bimodal Sensing System for Rehabilitation Electromechanical Device

Peng Wang, Shijie Guo

Biomedical Data

🎯 What it does: Developed a wearable mechanical pressure-electrophysiological dual-mode sensing system for monitoring in human limb rehabilitation devices.

Absolute Pose Estimation for a Millimeter-Scale Vision System

D. Ozturk, E. F. Helbling

Pose Estimation

🎯 What it does: Developed a millimeter-level visual system for achieving absolute pose estimation at 16.5 frames per second.