IROS 2024 Papers — Page 10
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers
Meta-Learning for Fast Adaptation in Intent Inferral on a Robotic Hand Orthosis for Stroke
Pedro Leandro La Rotta, M. Ciocarlie
Robotic IntelligenceMeta LearningSupervised Fine-TuningBiomedical Data
🎯 What it does: Propose MetaEMG, a meta-learning method for intention inference in robotic hand exoskeletons, which enables rapid adaptation to new sessions or new subjects with stroke patients.
METAVerse: Meta-Learning Traversability Cost Map for Off-Road Navigation
Junwon Seo, K. Kwak
Autonomous DrivingMeta LearningPoint Cloud
🎯 What it does: This paper proposes the METAVerse framework, which trains a global model using meta-learning to generate continuous traversal cost maps from sparse LiDAR point clouds. During deployment, it rapidly adapts to local environments through online adaptation. By integrating with a model predictive controller, it achieves safe and stable navigation on unknown irregular terrains.
MFC-EQ: Mean-Field Control with Envelope Q-learning for Moving Decentralized Agents in Formation
Qiushi Lin, Hang Ma
OptimizationReinforcement Learning
🎯 What it does: Propose the MFC-EQ framework, which utilizes mean field theory and Envelope Q-learning to address the dual-objective problem of maintaining desired formation and rapidly reaching the target in decentralized multi-agent path planning.
MFCalib: Single-shot and Automatic Extrinsic Calibration for LiDAR and Camera in Targetless Environments Based on Multi-Feature Edge
Tianyong Ye, Yukang Cui
Pose EstimationAutonomous DrivingImagePoint Cloud
🎯 What it does: Developed an MFCalib algorithm to achieve single automatic extrinsic calibration between LiDAR and RGB camera in a target-free environment.
MG-VLN: Benchmarking Multi-Goal and Long-Horizon Vision-Language Navigation with Language Enhanced Memory Map
Junbo Zhang, Kaisheng Ma
Autonomous DrivingVision Language ModelVision-Language-Action ModelSimultaneous Localization and MappingImageTextMultimodalityBenchmark
🎯 What it does: Proposed the MG-VLN benchmark task and studied the role of long-term memory and knowledge types in multi-goal long-range vision-language navigation.
MICP-L: Mesh-based ICP for Robot Localization Using Hardware-Accelerated Ray Casting
Alexander Mock, J. Hertzberg
Autonomous DrivingRobotic IntelligenceSimultaneous Localization and MappingMeshAgriculture Related
🎯 What it does: Propose a method called Mesh ICP Localization (MICP-L) for registering multiple ranging sensors with a triangular mesh map to achieve 6D real-time robot localization in GPS-denied environments.
Millipede-Inspired Multi-legged Magnetic Soft Robots for Targeted Locomotion in Tortuous Environments
Yibin Wang, Jiangfan Yu
Robotic Intelligence
🎯 What it does: Designed and tested a magnetic multi-legged soft robot with a zigzag flexible structure, capable of crawling on planes and slopes, and achieving precise navigation through magnetic field control.
Mind the Error! Detection and Localization of Instruction Errors in Vision-and-Language Navigation
Francesco Taioli, Yiming Wang
Anomaly DetectionAutonomous DrivingTransformerVision Language ModelMultimodalityBenchmark
🎯 What it does: Proposes a VLN-CE benchmark dataset containing various instruction errors and defines the task of instruction error detection and localization.
mini-PointNetPlus: A Local Feature Descriptor in Deep Learning Model for Real-time 3D Environment Perception
Chuanyu Luo, Pu Li
Autonomous DrivingRepresentation LearningConvolutional Neural NetworkPoint Cloud
🎯 What it does: Propose mini-PointNetPlus local feature descriptor for 3D perception, addressing the feature information loss in PointNet under max-pooling.
Miniaturisation and Evaluation of the SoftSCREEN System in Colon Phantoms
Vanni Consumi, A. Stilli
Robotic Intelligence
🎯 What it does: Design and miniaturize the SoftSCREEN system, conduct multiple tests on a colon model to assess the impact of pressure regulation on mobility
Miscommunication between robots can improve group accuracy in best-of-n decision-making
Raina Zakir, A. Reina
Robotic Intelligence
🎯 What it does: Investigated the impact of miscommunication on the speed-accuracy trade-off in robot collective decision-making, and employed a cross-inhibition model for binary classification decisions.
Mitigating Adversarial Perturbations for Deep Reinforcement Learning via Vector Quantization
T. Luu, C. D. Yoo
Adversarial AttackReinforcement LearningAuto Encoder
🎯 What it does: Propose using variant vector quantization (VQ) as a transformation method for input observations to reduce the adversarial attack space during the testing phase, thereby mitigating the impact of adversarial perturbations on reinforcement learning agents.
MIXED-SENSE: A Mixed Reality Sensor Emulation Framework for Test and Evaluation of UAVs Against False Data Injection Attacks
K. Pant, Inseok Hwang
Data SynthesisAnomaly DetectionRobotic Intelligence
🎯 What it does: Developed a high-fidelity hybrid reality sensor simulation framework for testing and evaluating the resilience of drones against false data injection attacks.
MLPER: Multi-Level Prompts for Adaptively Enhancing Vision-Language Emotion Recognition
Yu Gao, Honghai Liu
RecognitionLarge Language ModelPrompt EngineeringVision Language ModelContrastive LearningImageTextMultimodality
🎯 What it does: Proposed the MLPER model, which utilizes multi-level prompts (facial expressions, human posture, and scene conditions) through vision-language models for emotion recognition.
MM-Gaussian: 3D Gaussian-based Multi-modal Fusion for Localization and Reconstruction in Unbounded Scenes
Chenyang Wu, Yanyong Zhang
Gaussian SplattingSimultaneous Localization and MappingMultimodalityPoint Cloud
🎯 What it does: Propose the MM-Gaussian system, which utilizes LiDAR-camera multi-modal fusion to achieve localization and mapping in unmanned environments.
MM3DGS SLAM: Multi-modal 3D Gaussian Splatting for SLAM Using Vision, Depth, and Inertial Measurements
Lisong C. Sun, U. Topcu
Depth EstimationGaussian SplattingSimultaneous Localization and MappingImageMultimodalityPoint CloudTime Series
🎯 What it does: Construct a map using 3D Gaussian distributions and integrate visual, depth, and inertial measurements to achieve a precise SLAM system.
Mobility Performance Characterization of Transformable Nano Rover for Lunar Exploration
M. Sutoh, H. Sawada
Robotic IntelligencePhysics Related
🎯 What it does: Numerical simulation and ground test analysis were conducted on the mobility performance of the deformable nano-robot LEV-2 for lunar exploration.
ModaLink: Unifying Modalities for Efficient Image-to-PointCloud Place Recognition
Wei-Chau Xie, Xieyuanli Chen
RetrievalAutonomous DrivingImageMultimodalityPoint Cloud
🎯 What it does: Propose the ModaLink framework, which uses the FoV conversion module and non-negative decomposition encoder to unify images and point clouds into position-discriminative features
Model Agnostic Defense against Adversarial Patch Attacks on Object Detection in Unmanned Aerial Vehicles
Saurabh Pathak, Abdelrahman AlMahmoud
Object DetectionAdversarial AttackImage
🎯 What it does: Proposed a model-agnostic defense mechanism against adversarial patch attacks in UAV target detection
Model Predictive Control for Frenet-Cartesian Trajectory Tracking of a Tricycle Kinematic Automated Guided Vehicle
A. Subash, Karsten Bohlmann
Autonomous DrivingOptimization
🎯 What it does: Proposes an optimal control scheme for trajectory tracking of automated guided vehicles (AGVs) in warehouse environments, considering motion and collision constraints.
Model Predictive Path Integral Control for Agile Unmanned Aerial Vehicles
Michal Minarík, M. Saska
OptimizationComputational Efficiency
🎯 What it does: Proposed a real-time and onboard UAV control architecture that utilizes the MPPI method to achieve reference trajectory tracking and obstacle avoidance in cluttered environments.
Model Predictive Trees: Sample-Efficient Receding Horizon Planning with Reusable Tree Search
John Lathrop, Soon-Jo Chung
Autonomous DrivingOptimization
🎯 What it does: Proposed a recursive vision tree search algorithm for Model Predictive Trees (MPT), leveraging the reuse of entire optimal subtrees to enhance planning efficiency and quality.
Model-based Policy Optimization using Symbolic World Model
Andrey Gorodetskiy, Aleksandr Panov
OptimizationReinforcement LearningWorld Model
🎯 What it does: Propose using symbolic regression to generate symbolic expressions for approximating transition dynamics, leveraging symbolic dynamics models in model-based reinforcement learning to generate trajectories for improving sample efficiency, and evaluate in simulation tasks.
Modeling and Analysis of Passive Quadruped Walker with Compliant Torso on Low-friction Surface
Yuxuan Xiang, Fumihiko Asano
Robotic IntelligencePhysics Related
🎯 What it does: Modeling and analysis of passive quadruped walkers with compliant torsos on low-friction surfaces, investigating their walking stability and gait convergence under different friction coefficients, and observing performance trends through numerical simulation.
Modeling and Gait Analysis of Passive Rimless Wheel with Compliant Feet
Yanqiu Zheng, Isao T. Tokuda
Robotic IntelligencePhysics RelatedOrdinary Differential Equation
🎯 What it does: A passive elastic foot suitable for edgeless wheels was studied, and passive walking was achieved through numerical simulation, followed by analysis of elasticity and multi-cycle gait patterns.
Modeling of Hydraulic Soft Hand with Rubber Sheet Reservoir and Evaluation of its Grasping Flexibility and Control
Kyosuke Ishibashi, Ko Yamamoto
Robotic Intelligence
🎯 What it does: In this study, the authors derived the relationship between driving pressure, bending angle, and gripping force in a soft hand equipped with an elastic rubber membrane reservoir, and experimentally validated the soft hand's gripping flexibility when achieving angle control using this model.
Modelling and Analysis of Joint-to-End Variable Stiffness for Cable-Driven Hyper-Redundant Manipulator
Hongyang Zhang, Yuanlong Xie
Robotic IntelligencePhysics Related
🎯 What it does: A multi-layer static model was constructed by combining the virtual work principle with robotic arm kinematics. An analytical model of variable stiffness from joints to the end-effector and corresponding stiffness evaluation metrics were proposed. Subsequently, the relationship between cable tension, joint angles, joint stiffness, and end-effector stiffness was analyzed through platform validation and numerical methods.
Modernising Delivery: A Low-Energy Tethered Package System Using Fixed-Wing Drones
Samuel Ord, Timothy Wiley
Sequential
🎯 What it does: Proposed and verified a system using fixed-wing drones to deliver packages remotely through a long cable, designed the Mid-Tether Drag Device (MTDD) to improve the stability of the cable and package, provided the corresponding mathematical model, and conducted real flight verification under low wind conditions and on a flight field.
Modular Meshed Ultra-Wideband Aided Inertial Navigation with Robust Anchor Calibration
R. Jung, Stephan Weiss
Anomaly DetectionOptimizationComputational EfficiencySimultaneous Localization and Mapping
🎯 What it does: This paper proposes a generic filtering-based state estimation framework that supports two state decoupling strategies based on cross-covariance decomposition, and provides an autonomous calibration method for newly detected fixed devices (anchors) by combining multilabel and known anchor ranging data, along with an outlier rejection mechanism.
Modular Robot Wear for Walking Assistance According to Physical Functionality
Kunihiro Ogata, Yoshio Matsumoto
Robotic IntelligenceTime Series
🎯 What it does: A modular robotic wearable device was developed that can be customized according to user needs for gait assistance, supporting planar movements in the front-back and side directions.
MOE: A Dense LiDAR MOving Event Dataset, Detection Benchmark and LeaderBoard
Zhiming Chen, Hongyu Yu
Point CloudBenchmark
🎯 What it does: Constructed a LiDAR moving event dataset containing multi-scenario, high-density moving objects, and conducted a review and benchmark evaluation of existing moving event detection technologies.
Momentum-Aware Trajectory Optimisation using Full-Centroidal Dynamics and Implicit Inverse Kinematics
Aristotelis Papatheodorou, Ioannis Havoutis
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented a task-space optimization framework that directly optimizes footholds and contact forces for the quadruped robot ANYmal C using full centroidal dynamics to generate high-acceleration acrobatic motions.
Monocular 3D Reconstruction of Cheetahs in the Wild*
Zico da Silva, Amir Patel
Pose EstimationVideo
🎯 What it does: Proposed a monocular camera-based 3D reconstruction framework for cheetahs, integrating data-driven and physics-driven modeling with trajectory optimization.
Monocular Depth Estimation for Drone Obstacle Avoidance in Indoor Environments
Hao Zheng, A. Zakhor
Depth EstimationRobotic IntelligenceConvolutional Neural NetworkImage
🎯 What it does: Proposes a method for indoor obstacle avoidance and waypoint navigation based on monocular depth estimation, implemented on the 33g Bitcraze Crazyflie 2.1;
Monocular Event-Inertial Odometry with Adaptive decay-based Time Surface and Polarity-aware Tracking
Kai Tang, Jiajun Lv
Simultaneous Localization and Mapping
🎯 What it does: Proposes a monocular event-inertial odometry method that combines adaptive decay kernel time surfaces with polarity-aware tracking;
MonoForce: Self-supervised Learning of Physics-informed Model for Predicting Robot-terrain Interaction
R. Agishev, Tomáš Svoboda
Robotic IntelligenceImagePhysics Related
🎯 What it does: Developed an interpretable, physics-aware, end-to-end differentiable model that uses camera images to predict the results of robot-terrain interactions, applicable to both rigid and non-rigid terrains.
MonoPlane: Exploiting Monocular Geometric Cues for Generalizable 3D Plane Reconstruction
Wang Zhao, Hengkai Guo
Depth EstimationConvolutional Neural NetworkGraph Neural NetworkSupervised Fine-TuningImage
🎯 What it does: Proposes a general 3D plane detection and reconstruction framework called MonoPlane based on monocular geometric cues. It leverages pre-trained networks to extract depth and normals from single images, progressively fits planes using neighborhood-guided RANSAC, performs multi-plane joint optimization at the image level, and further extends to sparse view reconstruction.
MOSFormer: A Transformer-based Multi-Modal Fusion Network for Moving Object Segmentation
Zike Cheng, Ming Yang
SegmentationAutonomous DrivingTransformerMultimodality
🎯 What it does: Proposed the MOSFormer dual-branch multimodal fusion network and constructed the nuScenes-based MOS dataset for 3D moving object segmentation.
Motion Planning for Automata-based Objectives using Efficient Gradient-based Methods
Anand Balakrishnan, Jyotirmoy V. Deshmukh
Autonomous DrivingOptimization
🎯 What it does: Studies a motion planning method that utilizes gradient optimization to achieve symbolic automaton objectives
Motion Planning for Object Manipulation by Edge-Rolling
Maede Boroji, A. Fakhari
OptimizationRobotic Intelligence
🎯 What it does: This paper proposes a pre-grasping edge rolling motion approximation method based on rod theory, and designs a task space path generation algorithm utilizing rolling and pivot motion, to manipulate objects between two configurations.
Motion Primitives Planning For Center-Articulated Vehicles
Jiangpeng Hu, Marco Hutter
Autonomous DrivingAgriculture Related
🎯 What it does: Proposes a motion primitive planning method for Center Articulated Vehicles (CAV), integrating a recursive visibility planning framework with onboard sensors to achieve autonomous navigation.
MPGNet: Learning Move-Push-Grasping Synergy for Target-Oriented Grasping in Occluded Scenes
Dayou Li, Wei Zhang
Robotic Intelligence
🎯 What it does: Proposed the MPGNet three-branch network to learn the coordination of three actions (moving, pushing, grasping), and achieved target-oriented grasping through a multi-stage training strategy.
MPP: Multiscale Path Planning for UGV Navigation in Semi-structured Environments
Rui Cao, Wei Zhang
Autonomous DrivingOptimization
🎯 What it does: Propose a multi-scale path planning (MPP) method for unmanned ground vehicles (UGV) navigation in semi-structured environments, achieving path generation and obstacle handling through a three-layer architecture combining global planning, mid-level planning, and local trajectory planning.
MQE: Unleashing the Power of Interaction with Multi-agent Quadruped Environment
Ziyan Xiong, Yang Gao
Robotic IntelligenceReinforcement LearningAgentic AI
🎯 What it does: Proposed the Multi-Agent Quadruped Robot Environment (MQE), which provides multi-task collaboration and competition scenarios for the development and evaluation of multi-agent reinforcement learning algorithms.
MULAN-WC: Multi-Robot Localization Uncertainty-aware Active NeRF with Wireless Coordination
Weiying Wang, Stephanie Gil
Robotic IntelligenceNeural Radiance FieldSimultaneous Localization and Mapping
🎯 What it does: Proposes a multi-robot 3D reconstruction framework called MULAN-WC that utilizes wireless signal collaboration and NeRF.
Multi-Agent Behavior Retrieval: Retrieval-Augmented Policy Training for Cooperative Push Manipulation by Mobile Robots
So Kuroki, Tadashi Kozuno
Robotic IntelligenceTransformerReinforcement LearningRetrieval-Augmented Generation
🎯 What it does: Proposed a multi-agent coordination skill database and a retrieval-enhanced policy training method to achieve learning of collaborative pushing tasks with less data
Multi-agent Path Finding for Mixed Autonomy Traffic Coordination
Han Zheng, Cathy Wu
Autonomous DrivingReinforcement Learning
🎯 What it does: Proposed a hybrid automated traffic coordination behavior prediction motion priority search (BK-PBS) algorithm.
Multi-Agent Teamwise Cooperative Path Finding and Traffic Intersection Coordination
Zhongqiang Ren, Hesheng Wang
Autonomous DrivingOptimization
🎯 What it does: Studied the multi-agent team collaborative path planning (TCPF) problem and developed a centralized planner for signal-free intersection scenarios to generate collision-free paths and compute Pareto optimal solutions that balance travel time among teams.
Multi-agent Traffic Prediction via Denoised Endpoint Distribution
Yao Liu, Lina Yao
Autonomous DrivingTransformerDiffusion modelSequential
🎯 What it does: Proposed a denoising endpoint distribution model for trajectory prediction in high-speed scenarios
Multi-Agent Vulcan: An Information-Driven Multi-Agent Path Finding Approach
Jake Olkin, Brian C. Williams
Optimization
🎯 What it does: Proposed an information-driven multi-agent pathfinding method to address issues of redundant observations, collisions, and limited communication in multi-agent information gathering
Multi-Fidelity Reinforcement Learning for Minimum Energy Trajectory Planning
L. B. Castro, S. Karaman
Autonomous DrivingRobotic IntelligenceReinforcement Learning
🎯 What it does: A multi-fidelity Gaussian process model is used to predict quadrotor energy consumption by combining low-fidelity and high-fidelity samples, and this model is employed as a reward signal to drive reinforcement learning for generating minimum-energy trajectory planning.
Multi-Fingered Dragging of Unknown Objects and Orientations Using Distributed Tactile Information Through Vision-Transformer and LSTM
T. Ueno, S. Sugano
Robotic IntelligenceRecurrent Neural NetworkTransformer
🎯 What it does: Proposed a deep predictive learning method based on Vision-Transformer and LSTM for multi-fingered hand dragging operations on unknown objects with unknown poses
Multi-Fingered End-Effector Grasp Reflex Modeling for One-Shot Tactile Servoing in Tool Manipulation Tasks
E. Sheetz, Benjamin Kuipers
Robotic Intelligence
🎯 What it does: Proposed and implemented a grasp reflex model based on logistic regression for one-time tactile servoing in multi-fingered end-effectors during tool grasping tasks.
Multi-Fov-Constrained Trajectory Planning for Multirotor Safe Landing
Dong Wang, Fei Gao
Optimization
🎯 What it does: Propose a two-stage multi-FOV constrained trajectory planning algorithm. First, a safe initial path is generated through multi-FOV constrained path search, and then a safe landing trajectory satisfying FOV, dynamics, smoothness, and obstacle avoidance is generated via a safety landing trajectory optimization algorithm.
Multi-Goal Path Planning in Cluttered Environments with PRM-Guided Self-Organising Maps
Benjamin R. Davis, Graeme Best
OptimizationRobotic Intelligence
🎯 What it does: Research on the application of multi-robot multi-target path planning in crowded environments
Multi-modal Motion Prediction using Temporal Ensembling with Learning-based Aggregation
Kai Hong, Wen-Chieh Lin
Autonomous DrivingTransformerMultimodality
🎯 What it does: Propose a meta-algorithm based on time integration and learning aggregation to alleviate prediction inconsistency caused by missing behaviors in trajectory prediction.
Multi-modal NeRF Self-Supervision for LiDAR Semantic Segmentation
Xavier Timoneda, Fisher Yu
SegmentationNeural Radiance FieldImageMultimodalityPoint Cloud
🎯 What it does: Propose a semi-supervised learning method that utilizes an auxiliary NeRF head and the Segment-Anything (SAM) model from camera images to perform self-supervision on unlabeled LiDAR point clouds, thereby enhancing LiDAR semantic segmentation.
Multi-Modal Representation Learning with Tactile Data
Hyung-gun Chi, Thomas Kollar
Representation LearningRobotic IntelligenceContrastive LearningMultimodality
🎯 What it does: This paper proposes a multi-modal learning approach that combines tactile information with visual and language modalities, and constructs the Multi-Modal Wand (MMWand) dataset;
Multi-Robot Active Graph Exploration with Reduced Pose-SLAM Uncertainty via Submodular Optimization
Ruofei Bai, Lihua Xie
OptimizationRobotic IntelligenceSimultaneous Localization and MappingGraph
🎯 What it does: Propose a two-phase strategy that first generates a fast coverage path and then reduces the uncertainty in multi-robot collaborative SLAM by adding information-rich loop closure actions, modeling loop closure selection as a non-monotonic submodular maximization problem.
Multi-Robot Communication-Aware Cooperative Belief Space Planning with Inconsistent Beliefs: An Action-Consistent Approach
Tanmoy Kundu, Vadim Indelman
Robotic IntelligenceAgentic AI
🎯 What it does: Proposed a decentralized algorithm for multi-robot belief space planning to find consistent joint actions when robots hold inconsistent beliefs.
Multi-Robot Multi-Goal Mission Planning in Terrains of Varying Energy Consumption
Jáchym Herynek, Stefan Edelkamp
OptimizationRobotic Intelligence
🎯 What it does: Planning for multi-robot multi-task problems with energy constraints, considering dynamics, obstacles, collision avoidance, energy consumption, and charging requirements
Multi-Robot Navigation Among Movable Obstacles: Implicit Coordination to Deal with Conflicts and Deadlocks
Benoit Renault, Olivier Simonin
OptimizationRobotic Intelligence
🎯 What it does: This paper proposes a multi-robot version of the NAMO problem and studies conflict resolution and implicit coordination strategies.
Multi-Robot Path Planning With Boolean Specification Tasks Under Motion Uncertainties
Zhe Zhang, M. Reniers
Robotic IntelligenceReinforcement Learning
🎯 What it does: Study path planning for multiple robots to satisfy Boolean tasks under motion uncertainty, construct a global MDP, use local MDPs to compute locally optimal policies and update the global MDP, propose a heuristic reward function, and verify efficiency and scalability through numerical experiments.
Multi-Spectral Visual Servoing
Enrico Fiasché, Philippe Martinet
Robotic IntelligenceImage
🎯 What it does: A real-time visual servo method based on a multispectral camera was developed, utilizing dimensionality reduction techniques to extract pixels with the most informative features across all bands, while sacrificing spectral resolution to enhance spatial resolution;
Multi-Stage Monte Carlo Tree Search for Non-Monotone Object Rearrangement Planning in Narrow Confined Environments
Hanwen Ren, A. H. Qureshi
Robotic Intelligence
🎯 What it does: Propose a multi-stage Monte Carlo Tree Search (MS-MCTS) method for solving non-monotonic object rearrangement planning problems in narrow confined spaces, leveraging object phase topology to decompose complex problems into simpler subproblems, and designing a subgoal-focused tree expansion algorithm that simultaneously considers high-level planning and low-level robot motion.
Multi-target Tracking with Occlusion Resistance for Mobile Robots in Dynamic Environments*
Zhongyan Liu, Yongchun Fang
Object TrackingSegmentationRobotic IntelligenceTransformer
🎯 What it does: Propose a robust tracking algorithm called ROTrack for multi-target tracking of mobile robots in dynamic environments, which can accurately obtain the 3D coordinates of targets under occlusion conditions.
Multi-task real-robot data with gaze attention for dual-arm fine manipulation
Heecheol Kim, Y. Kuniyoshi
Robotic IntelligenceReinforcement LearningVision-Language-Action ModelImageTextMultimodality
🎯 What it does: Introduces a large-scale dataset containing dual-arm fine-grained operations (e.g., bowl relocation, pencil box opening, banana peeling) and proposes the Dual-Action and Attention model;
Multi-Uncertainty Aware Autonomous Cooperative Planning
Shiyao Zhang, Chengzhong Xu
Autonomous Driving
🎯 What it does: Propose a multi-uncertainty-aware autonomous cooperative planning (MUACP) framework that employs regularized cooperative model predictive control (RC-MPC) to simultaneously account for perception, motion, and communication uncertainties.
Multi-View 2D to 3D Lifting Video-Based Optimization: A Robust Approach for Human Pose Estimation with Occluded Joint Prediction*
Daniela Rato, Bogdan Raducanu
Pose EstimationOptimizationVideoBenchmark
🎯 What it does: Proposes an optimization method based on multi-view 2D-to-3D lifting for 3D human pose estimation in videos, incorporating temporal information.
Multidirectional slip detection and avoidance using dynamic 3D tactile meshes from visuotactile sensors
Peng Song, Y. Mezouar
Robotic IntelligenceConvolutional Neural NetworkImageMesh
🎯 What it does: Propose a new algorithm for multi-directional slip detection and avoidance using a dynamic 3D tactile mesh based on visual tactile sensors.
MultiGripperGrasp: A Dataset for Robotic Grasping from Parallel Jaw Grippers to Dexterous Hands
Luis Felipe Casas Murillo, Yu Xiang
Data SynthesisRobotic IntelligenceBenchmark
🎯 What it does: Constructed a large-scale grasping dataset called MultiGripperGrasp, containing 30.4 million grasps, covering 11 grippers and 345 objects, and validated grasp success and recorded drop times in the Isaac Sim simulator.
Multimodal Coherent Explanation Generation of Robot Failures
Pradip Pramanick, Silvia Rossi
ClassificationExplainability and InterpretabilityRobotic IntelligenceSupervised Fine-TuningMultimodality
🎯 What it does: Propose a method that generates coherent multimodal robotic fault explanations by checking the logical consistency between different modalities and performing necessary refinements.
Multimodal Evolutionary Encoder for Continuous Vision-Language Navigation
Zongtao He, Qi Chen
Autonomous DrivingVision-Language-Action ModelMultimodality
🎯 What it does: Proposed a multi-modal evolutionary encoder (MEE) that integrates multiple modalities such as depth and sub-instructions, employing evolutionary pre-training to enhance environmental understanding and generalization in continuous vision-language navigation.
Multimodal Failure Prediction for Vision-based Manipulation Tasks with Camera Faults
Yuliang Ma, Andrey Morozov
Anomaly DetectionRobotic IntelligenceVision-Language-Action ModelImageMultimodality
🎯 What it does: Proposed a multimodal fault prediction method for visual foundation manipulation systems, generated the FAULT-to-FAILURE dataset with 4000 real samples through fault injection, and subsequently trained the predictor using RGB images, mask images, and planned paths.
Multimodal Haptic Interface for Walker-Assisted Navigation
Yikun Wang, Carlos A. Cifuentes
Multimodality
🎯 What it does: Designed and evaluated a multimodal tactile handle integrating vibration, skin stretching, and combined feedback for walker navigation; compared the effectiveness of three feedback modes through experiments.
Multiple Visual Features in Topological Map for Vision-and-Language Navigation
Ruonan Liu, Weidong Zhang
Autonomous DrivingVision-Language-Action ModelSimultaneous Localization and MappingMultimodality
🎯 What it does: Proposed and implemented a new multi-visual feature topological map (MV-Topo) for audio-visual navigation in continuous environments.
MultipleCupSuctionNet: Deep Neural Network for Detecting Grasp Pose of a Vacuum Gripper with Multiple Suction Cups based on YOLO Feature Map Affine Transformation
Ping Jiang, Ooga Junichiro
Pose EstimationRobotic IntelligenceConvolutional Neural Network
🎯 What it does: Proposed and implemented a deep neural network called MultipleCupSuctionNet for detecting grasp poses of a multi-suction gripper.
Multistable Soft Actuator for Physical Human-robot Interaction
Juncai Long, Yixiong Feng
Robotic IntelligencePhysics Related
🎯 What it does: Designed a multi-stable soft actuator capable of memorizing arbitrary deformations without power consumption, and achieving multimodal deformation and tactile feedback through pneumatic bistable units.
MUP-LIO: Mapping Uncertainty-aware Point-wise Lidar Inertial Odometry
Hekai Yao, Zhuang Yan
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Propose a LiDAR-inertial odometry framework that combines uncertainty-aware mapping with point-to-plane matching and map refresh, addressing issues of point registration mismatch and LiDAR motion distortion.
MuTT: A Multimodal Trajectory Transformer for Robot Skills
Claudius Kienle, Gerhard Neumann
OptimizationRobotic IntelligenceTransformerMultimodality
🎯 What it does: Propose the Multimodal Trajectory Transformer (MuTT) architecture to integrate visual, trajectory, and robot skill parameters, predict robot skill execution under environmental perception, and combine with model-based optimizers to achieve skill parameter optimization without real-world execution.
MV-ROPE: Multi-view Constraints for Robust Category-level Object Pose and Size Estimation
Jiaqi Yang, L. Kneip
Pose EstimationSimultaneous Localization and MappingVideo
🎯 What it does: Proposes a multi-view method based on RGB video streams, combining scale-aware monocular dense SLAM, a lightweight object pose predictor, and an object-level pose graph optimizer to achieve category-level object pose and size estimation.
NanoNeRF: Robot-assisted Nanoscale 360° reconstruction with neural radiance field under scanning electron microscope
Xiang Fu, Song Liu
RestorationRobotic IntelligenceNeural Radiance FieldImage
🎯 What it does: Proposes a robot-assisted nanoscale 360° reconstruction method, which automatically captures multi-view images using a scanning electron microscope (SEM) and performs pixel-level structural reconstruction via neural radiance fields (NeRF).
NARRATE: Versatile Language Architecture for Optimal Control in Robotics
Seif Ismail, Carmen Amo Alonso
OptimizationTransformerLarge Language ModelText
🎯 What it does: Combined large language models (LLMs) with model predictive control (MPC) to achieve accurate, safe, and adjustable robot motion control through natural language.
Navigated Locomotion and Controllable Splitting of a Microswarm in a Complex Environment
Yuezhen Liu, Jiangfan Yu
Robotic Intelligence
🎯 What it does: Proposed a control strategy enabling the rope-like microrobot to achieve stable mode navigation and controllable splitting into two subgroups to reach two targets simultaneously.
NDT-Map-Code: A 3D global descriptor for real-time loop closure detection in lidar SLAM
L. Liao, Chun-yan Fu
Autonomous DrivingRepresentation LearningSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposed a global descriptor called NDT-Map-Code based on NDT for real-time loop detection in LiDAR SLAM;
Neighborhood Consensus Guided Matching Based Place Recognition with Spatial-Channel Embedding
Kun Li, Wei Liu
RecognitionImage
🎯 What it does: Proposes a spatial channel embedding (SCE) module and a neighborhood consensus guided matching (NCGM) module for visual place recognition, achieving more robust and discriminative global feature aggregation.
NeRF-enabled Analysis-Through-Synthesis for ISAR Imaging of Small Everyday Objects with Sparse and Noisy UWB Radar Data
Md. Farhan Tasnim Oshim, Tauhidur Rahman
RestorationSuper ResolutionNeural Radiance FieldImage
🎯 What it does: Propose a NeRF-based Analysis-through-Synthesis (ATS) framework to achieve high-resolution coherent ISAR imaging of small daily objects using sparse and noisy ultra-wideband (UWB) radar data.
Nerve Block Target Localization and Needle Guidance for Autonomous Robotic Ultrasound Guided Regional Anesthesia
Abhishek Tyagi, A. Trikha
Robotic IntelligenceConvolutional Neural NetworkSupervised Fine-TuningImageVideoBiomedical DataUltrasound
🎯 What it does: Developed a neural segmentation model for ultrasound-guided regional anesthesia, automated target localization (via ellipse fitting), and needle tip segmentation with trajectory extrapolation techniques, achieving a needle tip-guided autonomous navigation system.
NeSyMoF: A Neuro-Symbolic Model for Motion Forecasting
Achref Doula, Alejandro Sánchez Guinea
Autonomous DrivingExplainability and InterpretabilitySequentialBenchmark
🎯 What it does: Proposed a NeSyMoF model that combines deep neural networks with symbolic reasoning for motion prediction, capable of extracting features from the environment and generating first-order logic rules to explain the prediction process.
NeuFlow: Real-time, High-accuracy Optical Flow Estimation on Robots Using Edge Devices
Zhiyong Zhang, H. Singh
Robotic IntelligenceConvolutional Neural NetworkOptical Flow
🎯 What it does: Propose the NeuFlow architecture to achieve real-time high-precision optical flow estimation.
Neural Control Barrier Functions for Safe Navigation
Marvin Harms, Kostas Alexis
Autonomous DrivingOptimizationSafty and PrivacyRobotic IntelligencePoint Cloud
🎯 What it does: Developed a method that combines data learning with control barrier functions (CBF) and a safety controller, achieving real-time safe commands without maps or position estimation through a safety filter, and integrated LiDAR on a multirotor platform for simulation and field experiments.
Neural Kinodynamic Planning: Learning for KinoDynamic Tree Expansion
Tin Lai, Fabio Ramos
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented the L4KDE method, which uses a neural network to predict transfer costs between nodes to improve the kinodynamic tree expansion process.
Neural ODE-based Imitation Learning (NODE-IL): Data-Efficient Imitation Learning for Long-Horizon Multi-Skill Robot Manipulation
Shiyao Zhao, Zhibin Li
Robotic IntelligenceOrdinary Differential Equation
🎯 What it does: Developed a neural ODE-based imitation learning framework for efficiently learning multi-skill long-horizon robotic manipulation tasks
Neural Semantic Map-Learning for Autonomous Vehicles
Markus Herb, Federico Tombari
Autonomous DrivingSimultaneous Localization and Mapping
🎯 What it does: Proposes an integrated system that can fuse local submaps from a fleet into a coherent 3D grid map, encompassing elements such as drivable areas, lane markings, poles, and obstacles.
Neural Trajectory Model: Implicit Neural Trajectory Representation for Trajectories Generation
Zihan Yu, Yuqing Tang
Autonomous DrivingOptimizationComputational Efficiency
🎯 What it does: Reformulate single-agent and multi-agent trajectory planning problems as query problems for implicit neural trajectory representations, and propose Neural Trajectory Models (NTM) to generate near-optimal trajectories.
NeuralFloors++: Consistent Street-Level Scene Generation From BEV Semantic Maps
Valentina Muşat, Paul Newman
GenerationData SynthesisAutonomous DrivingDiffusion modelImage
🎯 What it does: Propose a two-stage generative model based on BEV semantic, instance, and style maps to synthesize photorealistic urban driving scenes and provide semantic, instance, and depth ground truth.
NeuralLabeling: A versatile toolset for labeling vision datasets using Neural Radiance Fields
Floris Erich, Y. Domae
Object DetectionSegmentationPose EstimationDepth EstimationSupervised Fine-TuningNeural Radiance FieldImage
🎯 What it does: Proposed a tool called NeuralLabeling that uses NeRF to annotate 3D scenes with bounding boxes or grids, generating segmentation masks, manipulability maps, 2D/3D bounding boxes, 6DOF pose, depth maps, and object meshes.
Neuro-Explorer: Efficient and Scalable Exploration Planning via Learned Frontier Regions
Kyung Min Han, Young J. Kim
OptimizationRobotic IntelligenceSimultaneous Localization and MappingBenchmark
🎯 What it does: Propose a learning-based, efficient, and scalable autonomous exploration system that utilizes three network models to identify frontier regions, assess their distance from the robot, and measure coverage rewards, achieving scalability through an active map window that moves with the robot;
Neuromorphic force-control in an industrial task: validating energy and latency benefits
Camilo Amaya, A. V. Arnim
Computational EfficiencyRobotic IntelligenceSpiking Neural NetworkReinforcement Learning
🎯 What it does: Trained a spiking neural network (SNN) for force-torque feedback control and migrated it to Intel Loihi neuromorphic chip, collaborating with KUKA robotic arm to accomplish industrial-level object insertion tasks
NeuSurfEmb: A Complete Pipeline for Dense Correspondence-based 6D Object Pose Estimation without CAD Models
Francesco Milano, Lionel Ott
Pose EstimationNeural Radiance FieldImage
🎯 What it does: Propose a complete 6D pose estimation pipeline trained with a small number of real images without requiring CAD models.