IROS 2024 Papers — Page 11
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers
NF-SLAM: Effective, Normalizing Flow-supported Neural Field representations for object-level visual SLAM in automotive applications
Li Cui, Zhenghua Yu
Autonomous DrivingOptimizationComputational EfficiencyRepresentation LearningFlow-based ModelNeural Radiance FieldSimultaneous Localization and MappingImagePoint Cloud
🎯 What it does: Proposes a vision-only object-level SLAM framework based on neural fields, utilizing implicit signed distance functions (SDF) combined with normalized flow networks to efficiently represent the 3D shapes of road vehicles.
NFPDE: Normalizing Flow-based Parameter Distribution Estimation for Offline Adaptive Domain Randomization
Rin Takano, Hiroyuki Oyama
Domain AdaptationFlow-based Model
🎯 What it does: Proposed a parameter distribution estimation method called NFPDE based on normalized flows, and validated its effectiveness in the OpenAI Gym environment.
NLNS-MASPF for solving Multi-Agent scheduling and Path-Finding
Heemang Park, Jinkyoo Park
OptimizationGraph Neural Network
🎯 What it does: Propose the NLNS-MASPF method to solve the multi-agent scheduling and pathfinding problem
Non-Repetitive: A Promising LiDAR Scanning Pattern
Angchen Xie, Ming Yang
Object DetectionPoint CloudBenchmark
🎯 What it does: Created the 'Repetitive-or-not' dataset containing both repetitive and non-repetitive LiDAR data collected simultaneously, conducted a comprehensive statistical analysis of the scanning capabilities of the two types of LiDAR, evaluated their impact on multiple 3D object detection algorithms, and explored the domain differences in point cloud data generated by the two LiDAR types.
Novel design of Reconfigurable Tracked Robot with Geometry-Changing Tracks
Chice Xuany, Yanjun Cao
Robotic Intelligence
🎯 What it does: Designed a robot with a reconfigurable geometry-adjustable track system, achieving good off-road performance on various terrains.
Novel Multiport Output Twisted String Actuator with Self-differential Mechanism: Hand Glove Application
Dunwen Wei, F. Ficuciello
Robotic Intelligence
🎯 What it does: Proposed a multi-port torsion rope actuator (MO-TSA) with a self-differential mechanism, and designed a glove capable of performing multiple grasping actions using a single actuator.
NRDF - Neural Region Descriptor Fields as Implicit ROI Representation for Robotic 3D Surface Processing
A. Pratheepkumar, Markus Vincze
Robotic IntelligenceNeural Radiance Field
🎯 What it does: Proposed the Neural Region Descriptor Field (NRDF) to achieve unsupervised dense 3D surface region correspondence, enabling retrieval of arbitrary processing-related regions of interest (P-ROI) in new instances of known categories, and applying it for one-click P-ROI-level process knowledge transfer.
NVINS: Robust Visual Inertial Navigation Fused with NeRF-augmented Camera Pose Regressor and Uncertainty Quantification
Juyeop Han, S. Karaman
Pose EstimationAutonomous DrivingRobotic IntelligenceNeural Radiance FieldImage
🎯 What it does: Propose a framework that fuses localization information generated by NeRF with Visual-Inertial Odometry (VIO), providing a robust solution for real-time robot navigation by training an absolute pose regression network based on NeRF-enhanced images and quantifying its uncertainty.
OAS-GPUCB: On-the-way Adaptive Sampling Using GPUCB for Bathymetry Mapping
Rajat Agrawal, P. B. Sujit
Autonomous DrivingOptimization
🎯 What it does: Proposed an OAS-GPUCB method based on GPUCB for adaptive sampling on lakes, aiming to minimize depth map errors while reducing measurement path length.
OBHMR: Robust Partial-to-full Generalized Point Set Registration with Overlap-guided Bidirectional Hybrid Mixture Model
Xinzhe Du, Zhe Min
Pose EstimationOptimizationPoint CloudBiomedical Data
🎯 What it does: Proposed an overlap-guided bidirectional hybrid model point set registration method (OBHMR), achieving robust partial-to-global geometric registration;
Object Augmentation Algorithm: Computing virtual object motion and object induced interaction wrench from optical markers
Christopher Herneth, Sami Haddadin
Robotic IntelligencePoint Cloud
🎯 What it does: Proposed and implemented an Object Augmentation Algorithm (OAA) that utilizes optical markers to calculate the hand coordinate system, virtual markers, inverse kinematics (IK), and inverse dynamics (ID), ultimately deriving virtual object motion and the induced joint torques, thereby expanding existing marker-based databases.
Object Instance Retrieval in Assistive Robotics: Leveraging Fine-Tuned SimSiam with Multi-View Images Based on 3D Semantic Map
Taichi Sakaguchi, T. Taniguchi
RetrievalRobotic IntelligenceContrastive LearningImage
🎯 What it does: Utilize multi-view images based on 3D semantic maps and adopt SimSiam self-supervised learning to train instance recognition models on-site, enabling instance-specific image goal navigation tasks in robots.
Object Pose Estimation by Camera Arm Control Based on the Next Viewpoint Estimation
Tomoki Mizuno, Tsuyoshi Tasaki
Pose EstimationRobotic IntelligenceImagePoint Cloud
🎯 What it does: Developed a neural network that simultaneously estimates pose and the next viewpoint for retail display robots handling products with simple shapes.
Object Segmentation from Open-Vocabulary Manipulation Instructions Based on Optimal Transport Polygon Matching with Multimodal Foundation Models
T. Nishimura, Komei Sugiura
SegmentationOptimizationVision-Language-Action ModelMultimodality
🎯 What it does: Open-vocabulary object manipulation instruction-based object segmentation mask generation
Object-based SLAM Using Superquadrics
Yifan Xing, Andrew Calway
Autonomous DrivingOptimizationSimultaneous Localization and Mapping
🎯 What it does: Proposes a SLAM system that uses superquadrics for object-level mapping and achieves camera tracking through keyframe optimization;
Object-Oriented Material Classification and 3D Clustering for Improved Semantic Perception and Mapping in Mobile Robots
Siva Krishna Ravipati, Suchendra Bhandarkar
ClassificationRobotic IntelligenceSimultaneous Localization and MappingImagePoint Cloud
🎯 What it does: Propose an object-oriented RGB-D material classification method and integrate it with ORB-SLAM2 to achieve 3D semantic mapping and multi-scale clustering
Occlusion Handling by Pushing for Enhanced Fruit Detection
Ege Gursoy, Andrea Cherubini
Object DetectionDepth EstimationRobotic IntelligenceGenerative Adversarial NetworkImageAgriculture Related
🎯 What it does: Detect occluded fruits using RGB-D cameras and robotic arms, estimate the occluded portions, determine the push-pull direction, and use the robotic arm to push away occluding branches to improve visibility.
ODD-diLLMma: Driving Automation System ODD Compliance Checking using LLMs
Carl Hildebrandt, Sebastian G. Elbaum
Autonomous DrivingTransformerLarge Language ModelText
🎯 What it does: Proposed the ODD-diLLMma method and evaluated it on three real-world datasets
ODTFormer: Efficient Obstacle Detection and Tracking with Stereo Cameras Based on Transformer
Tianye Ding, Huaizu Jiang
Object DetectionObject TrackingAutonomous DrivingTransformerImageBenchmark
🎯 What it does: Propose the ODTFormer model to achieve obstacle detection and tracking under stereo cameras;
Off-dynamics Conditional Diffusion Planners
Wen Zheng Terence Ng, Tianwei Zhang
OptimizationDiffusion modelScore-based ModelTime SeriesSequential
🎯 What it does: Propose using conditional diffusion probabilistic models (DPMs) to learn the joint distribution of large-scale offline dynamic datasets and limited target datasets, capturing environmental dynamics through two contexts (continuous dynamic scores and inverse dynamic contexts).
Offline Meta-Reinforcement Learning with Evolving Gradient Agreement
Jiaxing Chen, Peng Li
Meta LearningReinforcement Learning
🎯 What it does: Proposes a new offline Meta-reinforcement learning method called MACEGA that can quickly adapt to unseen tasks using offline data
OmniNxt: A Fully Open-source and Compact Aerial Robot with Omnidirectional Visual Perception
Peize Liu, Shaojie Shen
Robotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Proposed and implemented OmniNxt, a fully open-source lightweight aerial robot platform with omnidirectional visual perception, including a high-performance flight controller Nxt-FC and a multi-fisheye camera array, along with the development of compatible software for precise localization and real-time dense mapping.
OmniRace: 6D Hand Pose Estimation for Intuitive Guidance of Racing Drone
Valerii Serpiva, D. Tsetserukou
Pose EstimationConvolutional Neural Network
🎯 What it does: Propose the OmniRace technology, which uses 6-DoF gesture recognition to control racing drones.
On a Magnetically Driven Array System with Autonomous Motion and Object Delivery for Biomedical Microrobots
Yueyue Liu, Qigao Fan
Robotic IntelligenceDrug DiscoveryConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: Develop a PCB magnetic field-driven multi-micro-robot system based on micro-coil arrays, utilizing local magnetic fields to achieve independent control of multiple micro-robots, and integrating perception, planning, and execution for autonomous multi-task drug delivery.
On Learning Scene-aware Generative State Abstractions for Task-level Mobile Manipulation Planning
Julian Förster, Roland Siegwart
ClassificationGenerationRobotic IntelligenceGraph Neural NetworkGenerative Adversarial NetworkPoint Cloud
🎯 What it does: Learning a system that can both classify the states of predicates in scenes and generate scene configurations that satisfy specified predicates
On performing non-prehensile rolling manipulations: Stabilizing synchronous motions of Butterfly robots⋆
Maksim O. Surov, L. Freidovich
Robotic Intelligence
🎯 What it does: Designed a model-based centralized feedback controller to achieve synchronized rolling of spheres on multiple Butterfly robots.
On Predicting Terrain Changes Induced by Mobile Robot Traversal
Miloš Prágr, J. Faigl
Robotic IntelligenceConvolutional Neural Network
🎯 What it does: Predict terrain changes caused by robot movement and support subsequent vehicle decision-making.
On the 3D trochoidal motion model of LiDAR sensors placed off-centered inside spherical mobile mapping systems
Fabian Arzberger, Andreas Nüchter
Pose EstimationAutonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Studied the motion model of sensors fixed inside a spherical mobile mapping platform and proposed a new calibration method for spherical systems to estimate the extrinsic parameters of the sensors relative to the sphere's rotation center. Subsequently, the calibration and motion model were deployed on the platform to estimate the LiDAR sensor trajectory, and compared with advanced LiDAR-IMU odometry methods.
On the Benefits of GPU Sample-Based Stochastic Predictive Controllers for Legged Locomotion
Giulio Turrisi, Claudio Semini
Computational EfficiencyRobotic Intelligence
🎯 What it does: Proposed and implemented a GPU-accelerated sample-based stochastic predictive control method for gait frequency adaptation in quadruped robots, and compared it with traditional gradient-based model predictive control (MPC).
On the Modularity of Elementary Dynamic Actions
Moses C. Nah, Neville Hogan
Robotic Intelligence
🎯 What it does: Proposes a kinematic modular control method based on Elementary Dynamic Actions, which can generate diverse robot motions by combining basic modules.
One Problem, One Solution: Unifying Robot Design and Cell Layout Optimization
J. Baumgärtner, J. Fleischer
OptimizationRobotic Intelligence
🎯 What it does: Proposed a unified problem formulation for jointly optimizing robot kinematics and workstation layout, and validated its effectiveness on a robotic milling system.
One-Shot Transfer of Long-Horizon Extrinsic Manipulation Through Contact Retargeting
Albert Wu, C. K. Liu
Robotic Intelligence
🎯 What it does: Proposes a method through contact retargeting, which remaps contact requirements to generalize a single short-term primitive strategy library to various objects and environments, enabling long-horizon external manipulation tasks.
Online Adaptation of Learned Vehicle Dynamics Model with Meta-Learning Approach
Yuki Tsuchiya, Guy Rosman
Autonomous DrivingMeta LearningTime SeriesSequential
🎯 What it does: A vehicle dynamics model is constructed using a multi-layer neural network, and Continual-MAML is employed to achieve online adaptation, enabling the model to rapidly adapt to new environments without forgetting previous ones.
Online Adaptive Impedance Control with Gravity Compensation for an Interactive Lower-Limb Exoskeleton
Run Janna, P. Manoonpong
Robotic Intelligence
🎯 What it does: Studies an interactive lower-limb exoskeleton control method combining online adaptive impedance control and gravity compensation to achieve collaborative interaction between the user and the exoskeleton;
Online Determination of Legged Kinematics
Chinmay Burgul, Guoquan Huang
Robotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Online state estimation identifies leg-body kinematic parameters (leg-body transformation, time offset, leg segment length) for robots with an arbitrary number of legs
Online Efficient Safety-Critical Control for Mobile Robots in Unknown Dynamic Multi-Obstacle Environments
Yu Zhang, A. Knoll
Computational EfficiencyRobotic IntelligencePoint Cloud
🎯 What it does: Proposes a LiDAR-based target tracking and exploration framework for achieving online efficient safety-critical control in unknown dynamic multi-obstacle environments.
Online Hand Movement Recognition System with EEG-EMG Fusion Using One-Dimensional Convolutional Neural Network
Haozheng Wang, Feng Duan
ClassificationRecognitionConvolutional Neural NetworkTime SeriesBiomedical Data
🎯 What it does: A 1D-CNN-based EEG-EMG fusion hand motion recognition system was developed, classifying EEG and EMG data from five subjects in offline experiments, while the trained model was used for online identification to control the Pepper robot to perform corresponding hand motions.
Online Multi-Agent Pickup and Delivery with Task Deadlines
Hiroya Makino, Seigo Ito
Optimization
🎯 What it does: Proposes the deadline-driven problem for online multi-agent transportation and delivery tasks (MAPD-D), and designs deadline-aware token passing (D-TP) and task-swapping enhanced D-TP (D-TPTS) algorithms to handle task deadlines.
Online Optimization of Central Pattern Generators for Quadruped Locomotion
Zewei Zhang, A. Ijspeert
OptimizationRobotic Intelligence
🎯 What it does: Adapt quickly to speed commands and terrain changes (friction coefficient, slope, load) by online optimizing the central pattern generator (CPG) parameters of quadruped robots through Bayesian optimization.
Online Planning for Multi Agent Path Finding in Inaccurate Maps
Nir Malka, Roni Stern
Optimization
🎯 What it does: Propose a multi-agent path planning method that alternates between online planning and execution to handle inaccurate maps, reducing replanning overhead by replanning only for affected agents and postponing future conflicts, while adapting single-agent path planning algorithms.
Online Refractive Camera Model Calibration in Visual Inertial Odometry
Mohit Singh, Kostas Alexis
Autonomous DrivingOptimizationRobotic IntelligenceSimultaneous Localization and MappingVideoPhysics Related
🎯 What it does: Proposed a general refraction camera model and online co-estimated the refractive index and pose of unknown media within a monocular visual-inertial odometry framework, achieving operation in various refractive fluids with only air-based camera calibration.
Online Rotor Fault Detection and Isolation for Vertical Takeoff and Landing Vehicles
Jiaqi Lian, L. T. X. Phan
Anomaly DetectionTime Series
🎯 What it does: Online detection and isolation of rotor faults in vertical take-off and landing (VTOL) vehicles.
Online Tree Reconstruction and Forest Inventory on a Mobile Robotic System
Leonard Freißmuth, Maurice F. Fallon
Robotic IntelligenceSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Developed a real-time mapping and analysis system that uses a laser scanner mounted on a mobile robot or backpack to generate forest inventories online; the system is based on incrementally generated submaps, employing a custom Voronoi heuristic clustering to extract tree candidates, then using a Hough transform-based algorithm for robust trunk modeling, and continuously updating the database through pose graph LiDAR SLAM to refine tree feature estimates during subsequent revisits to the area.
Ontology Based AI Planning and Scheduling for Robotic Assembly
Jingyun Zhao, André Kraft
Robotic Intelligence
🎯 What it does: Propose a dynamic task planning and scheduling method based on ontology and artificial intelligence to enhance robot assembly efficiency, reduce downtime, and increase production capacity.
Open Human-Robot Collaboration using Decentralized Inverse Reinforcement Learning
Prasanth Sengadu Suresh, Diego Romeres
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed the oDec-MDP multi-agent framework for modeling open human-robot collaboration, and extended the Dec-AIRL method to this framework for inverse reinforcement learning.
Open6DOR: Benchmarking Open-instruction 6-DoF Object Rearrangement and A VLM-based Approach
Yufei Ding, He Wang
Robotic IntelligenceTransformerLarge Language ModelVision Language ModelBenchmark
🎯 What it does: Created the Open6DOR benchmark for 6-DoF object reordering with open instructions, and proposed the Open6DOR-GPT method based on GPT-4V
OPENGRASP-LITE Version 1.0: A Tactile Artificial Hand with a Compliant Linkage Mechanism
Sonja Groß, Sami Haddadin
Robotic Intelligence
🎯 What it does: Proposed an open-source, lightweight, and highly integrated tactile artificial hand with a flexible linkage mechanism enabling multi-functional grasping, equipped with six degrees of freedom actuation and MEMS tactile sensors on each fingertip.
OpenOcc: Open Vocabulary 3D Scene Reconstruction via Occupancy Representation
Haochen Jiang, Li Zhang
GenerationKnowledge DistillationRepresentation LearningNeural Radiance Field
🎯 What it does: Proposes the OpenOcc framework, unifying 3D scene reconstruction with open-vocabulary understanding through occupancy representations and volume rendering that distills pre-trained open-vocabulary models into 3D language fields.
OPG-Policy: Occluded Push-Grasp Policy Learning with Amodal Segmentation
Hao Ding, Hui Cheng
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposes the OPG-Policy framework, which predicts the occluded parts of objects in dense cluttered environments through full-modal segmentation and achieves grasping of partially occluded targets by combining push-pull-grasp action sequences
Opinion-based Strategy for Distributed Multi-Robot Task Allocation in Swarms of Robots
Ziqiao Zhang, Fumin Zhang
OptimizationRobotic Intelligence
🎯 What it does: Propose a distributed multi-robot task allocation strategy based on opinion dynamics
Opti-Acoustic Semantic SLAM with Unknown Objects in Underwater Environments
Kurran Singh, John J. Leonard
Simultaneous Localization and MappingMultimodality
🎯 What it does: Proposes an object-based underwater semantic SLAM method that can identify, locate, classify, and map various marine objects in environments with unknown object categories.
Optimal and Bounded Suboptimal Any-Angle Multi-agent Pathfinding
Konstantin S. Yakovlev, Roni Stern
Optimization
🎯 What it does: Proposed the first optimal arbitrary-angle multi-agent path planning algorithm.
Optimal Integration of Hybrid FES-Exoskeleton for Precise Knee Trajectory Control
Masoud Jafaripour, Mahdi Tavakoli
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented a novel hybrid torque allocation method to integrate quadriceps functional electrical stimulation (FES) with powered exoskeleton systems, enhancing wearability and mobility.
Optimal Robot Formations: Balancing Range-Based Observability and User-Defined Configurations
S. S. Ahmed, J. R. Forbes
OptimizationRobotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Introduce a customizable cost function to balance robot spacing, achieving high relative pose estimation accuracy and specific task requirements (such as high coverage), and evaluate the desired formation in coverage path planning.
Optimal Robotic Assembly Sequence Planning (ORASP): A Sequential Decision-Making Approach
Kartik Nagpal, Negar Mehr
Robotic IntelligenceReinforcement Learning
🎯 What it does: Propose a method that models robot assembly planning as a Markov decision process, and utilize dynamic programming, Graph Exploration Assembly Planner (GEAP), ORASP search, and deep reinforcement learning to generate optimal assembly sequences;
Optimal Sensing in Soft Pneumatic Actuators via Stretchable Optical Waveguides
Faisal Al Jaber, P. Choe
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Studied the impact of the position and layout of stretchable optical waveguide sensors in flexible pneumatic actuators on responsiveness, repeatability, and durability, conducting experiments with three different exoskeleton reinforcement configurations.
Optimal view point and kinematic control for grape stem detection and cutting with an in-hand camera robot
Sotiris Stavridis, Zoe Doulgeri
OptimizationRobotic IntelligencePoint CloudAgriculture Related
🎯 What it does: Propose a method for finding the optimal view and cutting angle for grape stems, and achieve control through classified point clouds obtained from a hand-held camera;
Optimizing Base Placement of Surgical Robot: Kinematics Data-Driven Approach by Analyzing Working Pattern
Jeonghyeon Yoon, Minho Hwang
OptimizationRobotic IntelligenceBiomedical Data
🎯 What it does: Propose a method for determining the optimal base posture of a surgical robot based on the working patterns of surgeons, utilizing machine learning clustering analysis to record end-effector postures, identifying key positions and orientations, and introducing joint margin and manipulability scores, using a multi-layer perceptron regressor to predict the optimal base posture.
Optimizing Crowd-Aware Multi-Agent Path Finding through Local Communication with Graph Neural Networks
Phu-Cuong Pham, Aniket Bera
OptimizationGraph Neural NetworkReinforcement Learning
🎯 What it does: Propose a decentralized reinforcement learning method called CRAMP based on graph neural networks to address multi-agent pathfinding (MAPF) problems in crowded environments;
Optimizing Interaction Space: Enlarging the Capture Volume for Multiple Portable Motion Capture Devices
M. Fatoni, Sami Haddadin
Optimization
🎯 What it does: Using four Leap Motion Controllers (LMC) to determine the optimal layout through Monte Carlo simulations, and evaluating its reliability and effectiveness with 1560 trials on 10 subjects compared to a marker-based motion capture system (MMC).
Optimizing Kubernetes Deployment of Robotic Applications with HEFT-based Container Orchestration
Francesco Lumpp, N. Bombieri
OptimizationRobotic Intelligence
🎯 What it does: Proposes HEFT4K, an event-driven scheduling method based on HEFT for robot software deployment in Kubernetes environments while satisfying QoS constraints.
Origami Actuator with Tunable Limiting Layer for Reconfigurable Soft Robotic Grasping
Yang Yang, Yingtian Li
Robotic Intelligence
🎯 What it does: Designed and fabricated a flexible origami actuator with a tunable limit layer for reconfigurable soft mechanical grasping
OSM vs HD Maps: Map Representations for Trajectory Prediction
Jing-Yan Liao, Henrik I. Christensen
Autonomous DrivingGraphSequential
🎯 What it does: This paper proposes using OpenStreetMap (OSM) as an alternative to high-definition maps (HD Map) for long-term motion prediction, enhancing prediction performance by expanding the observation horizon and incorporating intersection prior knowledge; meanwhile, a comprehensive context-aware analysis is conducted to deeply explore motion prediction characteristics across different scenarios and categories.
OTVIC: A Dataset with Online Transmission for Vehicle-to-Infrastructure Cooperative 3D Object Detection
He Zhu, Yue Wang
Object DetectionAutonomous DrivingTransformerMultimodality
🎯 What it does: Proposed the OTVIC multimodal multi-view dataset and the LfFormer framework based on Transformer for multimodal delayed fusion
Outlier-Robust Geometric Perception: A Novel Thresholding-Based Estimator with Intra-Class Variance Maximization
Lei Sun
Anomaly DetectionOptimization
🎯 What it does: Proposed a general robust estimator TIVM that can collaborate with standard non-minimal solvers to efficiently remove outliers in geometric perception problems.
OV-MAP : Open-Vocabulary Zero-Shot 3D Instance Segmentation Map for Robots
Juno Kim, Byoung-Tak Zhang
SegmentationRobotic IntelligencePoint Cloud
🎯 What it does: Propose a 3D mapping method utilizing open vocabulary features to achieve zero-shot 3D instance segmentation.
OVGNet: A Unified Visual-Linguistic Framework for Open-Vocabulary Robotic Grasping
Meng Li, Chenguang Yang
Robotic IntelligenceVision Language ModelVision-Language-Action ModelMultimodalityBenchmark
🎯 What it does: Proposed the OVGNet framework, integrating open-vocabulary learning to enable robots to grasp objects of new categories, and constructed a corresponding large-scale benchmark dataset.
OW3Det: Toward Open-World 3D Object Detection for Autonomous Driving
Wenfei Hu, Dingsheng Luo
Object DetectionAutonomous DrivingKnowledge DistillationPoint Cloud
🎯 What it does: Proposed the OW3Det framework for open-world 3D object detection, addressing closed-set prediction and incremental learning problems;
P4: Pruning and Prediction-based Priority Planning
Rui Yang, Rajiv Gupta
Optimization
🎯 What it does: Proposes the P4 algorithm, which integrates the point-to-point (PnP) algorithm, dynamic window approach, and path direction prediction to address multi-agent path planning problems.
PA-LOCO: Learning Perturbation-Adaptive Locomotion for Quadruped Robots
Zhiyuan Xiao, Qingrui Zhang
Robotic IntelligenceReinforcement Learning
🎯 What it does: Proposed and implemented a privilege learning framework incorporating multiple feature encoders and residual strategy networks to enhance the robust locomotion performance of quadruped robots across various terrains and external disturbances, validated on the Unitree GO1 robot.
PAAMP: Polytopic Action-Set And Motion Planning for Long Horizon Dynamic Motion Planning via Mixed Integer Linear Programming
A. Jaitly, Siavash Farzan
OptimizationRobotic Intelligence
🎯 What it does: Proposes the PAAMP method, which utilizes polygonal action sets to convert long-horizon dynamic motion planning problems into linear programming and further solves them as mixed-integer linear programming (MILP).
PACC: A Passive-Arm Approach for High-Payload Collaborative Carrying with Quadruped Robots Using Model Predictive Control
Giulio Turrisi, V. Barasuol
Robotic Intelligence
🎯 What it does: This paper proposes the use of passive arm structures with intrinsic impedance in quadruped robot collaborative transportation tasks, guided by arm joint displacement to direct robot movement; simultaneously, a decentralized model predictive controller (MPC) is designed to control robot locomotion and estimate external forces generated during collaborative transportation. The effectiveness of the system is verified through robot-robot and human-robot collaborative transportation experiments on stair obstacles and rugged terrains.
PanopticRecon: Leverage Open-vocabulary Instance Segmentation for Zero-shot Panoptic Reconstruction
Xuan Yu, Yue Wang
Image
🎯 What it does: Propose a zero-shot panoramic reconstruction method based on RGB-D images
Parametric Synthesis of Compliant Joints for Impact-Robust Shaftless Leg Mechanisms
E. Rakshin, S. Kolyubin
OptimizationRobotic IntelligencePhysics Related
🎯 What it does: Proposed and implemented parameterized optimization and experimental validation of a three-degree-of-freedom flexible cross hinge in a closed-loop leg mechanism.
ParametricNet++: A 6DoF Pose Estimation Network with Sparse Keypoint Recovery for Parametric Shapes in Stacked Scenarios
Yihan Xie, Long Zeng
Pose Estimation
🎯 What it does: Proposed the ParametricNet++ network for 6DoF pose estimation, combining point-to-point regression and sparse keypoint recovery, while enhancing performance through keypoint selection and prediction optimization.
PARE: A Plane-Assisted Autonomous Robot Exploration Framework in Unknown and Uneven Terrain
Pu Xu, Zheng Fang
Robotic Intelligence
🎯 What it does: Proposes a plane information-based autonomous robot exploration framework (PARE) to achieve maximum volume and safe autonomous exploration.
ParkingE2E: Camera-based End-to-end Parking Network, from Images to Planning
Changze Li, Ming Yang
Autonomous DrivingTransformerReinforcement LearningImage
🎯 What it does: Proposed an end-to-end parking network from RGB images to path planning, using imitation learning to learn and execute human driving trajectories
Passive Underwater Robot Hand Utilizing Water Resistance
Issei Nate, S. Hirai
Robotic Intelligence
🎯 What it does: Designed and experimentally verified a multi-fingered underwater robotic gripper that utilizes water resistance to achieve opening/closing and employs a fingertip locking mechanism for passive locking/unlocking.
Path Re-Planning with Stochastic Obstacle Modeling: A Monte Carlo Tree Search Approach
Francesco Trotti, Riccardo Muradore
Autonomous DrivingOptimizationReinforcement LearningBenchmark
🎯 What it does: Proposes a local path replanning strategy based on Monte Carlo Tree Search (MCTS) for unmanned ground vehicles (UGVs) to replan paths when encountering unexpected obstacles in dynamic warehouse environments.
Path-Parameterised RRTs for Underactuated Systems
Damian Abood, Ian R. Manchester
OptimizationRobotic Intelligence
🎯 What it does: Propose a sampling-based motion planning algorithm for underactuated systems based on path parameterization, and implement a dedicated state-driven guide mechanism within RRT to simultaneously generate geometric paths and their time parameterization.
PathFinder: Attention-Driven Dynamic Non-Line-of-Sight Tracking with a Mobile Robot
Shenbagaraj Kannapiran, Spring Berman
Object TrackingRobotic IntelligenceTransformerImageVideo
🎯 What it does: Propose a dynamic non-line-of-sight (NLOS) tracking method called PathFinder based on a standard RGB camera, using a neural network with an attention mechanism to estimate the 2D trajectory of hidden individuals in real-time on a mobile robot.
PathFormer: A Transformer-Based Framework for Vision-Centric Autonomous Navigation in Off-Road Environments
Bilal Hassan, Jorge Dias
Autonomous DrivingTransformerImage
🎯 What it does: Designed a Transformer-based end-to-end framework called PathFormer, which directly decodes free space semantics and configurations using camera images, achieving path planning without the need for pre-mapping.
Patterned Structure Muscle : Arbitrary Shaped Wire-driven Artificial Muscle Utilizing Anisotropic Flexible Structure for Musculoskeletal Robots
Shunnosuke Yoshimura, Masayuki Inaba
Robotic Intelligence
🎯 What it does: A Patterned Structure Muscle (PSM) was proposed, utilizing patterned structures with anisotropic characteristics and a wire-driven mechanism, fabricated using FDM 3D printing technology with TPU soft material, capable of fabricating artificial muscles of arbitrary shapes and demonstrating their functionality in an upper arm structure.
PCDepth: Pattern-based Complementary Learning for Monocular Depth Estimation by Best of Both Worlds
Haotian Liu, Guang Chen
Depth EstimationRepresentation LearningMultimodality
🎯 What it does: Proposed the PCDepth patterned complementary learning architecture for monocular depth estimation, which enhances depth prediction performance by discretizing scenes into high-level patterns and integrating across different modalities.
PCT: Perspective Cue Training Framework for Multi-Camera BEV Segmentation
Haruya Ishikawa, Yoshimitsu Aoki
SegmentationDomain AdaptationAutonomous DrivingImage
🎯 What it does: Propose the Perspective Cue Training (PCT) framework, utilizing pseudo-labels generated from unannotated perspective images to train BEV segmentation with a shared image encoder.
Pedicle Drilling Planning Transfer for Spine Surgery Using Functional Map Correspondences
L. Leblanc, B. Tamadazte
Robotic IntelligenceBiomedical DataComputed Tomography
🎯 What it does: Propose a method that utilizes the functional mapping (FM) framework to transfer needle paths from pre-operative CT scans to partially observed and noisy spinal models.
PEERNet: An End-to-End Profiling Tool for Real-Time Networked Robotic Systems
Aditya Narayanan, Sandeep P. Chinchali
Robotic Intelligence
🎯 What it does: Developed an end-to-end real-time performance analysis tool called PEERNet, specifically designed for monitoring and evaluation of cloud robotic systems.
PEGASUS: Physically Enhanced Gaussian Splatting Simulation System for 6DoF Object Pose Dataset Generation
Lukas Meyer, Y. Domae
Data SynthesisPose EstimationGaussian SplattingImagePhysics Related
🎯 What it does: Proposed and implemented PEGASUS, a physics-enhanced simulation system based on 3D Gaussian Splatting, to generate a synthetic dataset containing RGB images, depth maps, semantic masks, and 6DoF object poses, and verified its transferability on real data.
Perception for Connected Autonomous Vehicles under Adverse Weather Conditions
Dimitra Tsakmakopoulou, Konstantinos Moustakas
Object DetectionAutonomous DrivingConvolutional Neural NetworkPoint Cloud
🎯 What it does: The paper demonstrates the necessity of maintaining high detection accuracy through vehicle-to-vehicle (V2V) communication in adverse weather conditions, proposing and evaluating a collaborative perception system, particularly tested in foggy scenarios.
Perception-aware Full Body Trajectory Planning for Autonomous Systems using Motion Primitives
Moritz Kuhne, Máximo A. Roa
Autonomous DrivingRobotic Intelligence
🎯 What it does: Proposed a perception-aware trajectory planner for robot systems with independently adjustable camera directions, using motion primitives to achieve planning.
Perception-Driven Shared Control Architecture for Agricultural Robots Performing Harvesting Tasks
Jozsef Palmieri, Alessandro Marino
OptimizationRobotic IntelligenceAgriculture Related
🎯 What it does: Developed a dynamic switching human-machine shared control framework based on perceived uncertainty for operation of agricultural mobile machinery in harvesting tasks
Perfecting Periodic Trajectory Tracking: Model Predictive Control with a Periodic Observer (Π-MPC)
Luis A. Pabon, Marco Pavone
Autonomous DrivingOptimizationRobotic Intelligence
🎯 What it does: A simple controller combining an observer with a Model Predictive Control (MPC) scheme for periodic trajectory tracking was designed, achieving asymptotic perfect tracking even in the presence of model mismatch.
Performing Efficient and Safe Deformable Package Transport Operations Using Suction Cups
Rishabh Shukla, Satyandra K. Gupta
OptimizationRobotic Intelligence
🎯 What it does: This paper proposes a method to calculate the safety factor (FOS) for each trajectory point by analyzing the curvature of the contact interface between the package and the suction cup, and maintains the FOS threshold during trajectory planning to achieve safe and time-optimized transport of deformable packages.
Peristaltic Soft Robot for Long-distance Pipe Inspection with an Endoskeletal Structure for Propulsion and Traction Amplification
R. Okuma, T. Nakamura
Robotic Intelligence
🎯 What it does: Developed a peristaltic pipeline inspection robot using a linear antagonistic mechanism with artificial muscles and an internal skeleton structure to amplify thrust and traction for long-distance, narrow, and complex pipeline inspections.
PGA: Personalizing Grasping Agents with Single Human-Robot Interaction
Junghyun Kim, Byoung-Tak Zhang
Domain AdaptationRobotic IntelligenceSupervised Fine-TuningImage
🎯 What it does: Propose a robot grasping agent (PGA) that achieves personalized grasping through a single human-robot interaction and a pseudo-label propagation algorithm utilizing unlabeled image data.
PhotoBot: Reference-Guided Interactive Photography via Natural Language
Oliver Limoyo, Gregory Dudek
Object DetectionPose EstimationTransformerLarge Language ModelVision Language ModelImageTextRetrieval-Augmented Generation
🎯 What it does: Propose a fully automatic photography framework called PhotoBot, which is based on high-level human language guidance and interaction with a robot photographer, conveying photography suggestions to users by selecting reference images from a curated image set;
Photometric Consistency for Precise Drone Rephotography
Hsaun-Jui Chang, Kuan-Wen Chen
Object TrackingPose EstimationSimultaneous Localization and MappingOptical FlowImage
🎯 What it does: A precise UAV rephotography system for fixed-domain patrol scenarios is proposed, which combines computer vision-based localization and fine-grained pixel-level dense flow prediction, and introduces an interleaved UAV controller flight scheme;
Physically Consistent Online Inertial Adaptation for Humanoid Loco-manipulation
James Foster, Robert J. Griffin
Robotic IntelligencePhysics Related
🎯 What it does: Designed and implemented an online estimation and control framework that employs a physics-consistent extended Kalman filter (EKF) for inertial parameter estimation, integrated with a whole-body controller, enabling humanoid robots to perform motion and manipulation tasks under large external loads.
Physically-Based Photometric Bundle Adjustment in Non-Lambertian Environments
Lei Cheng, Haoang Li
Pose EstimationDepth EstimationSimultaneous Localization and MappingImagePoint CloudPhysics Related
🎯 What it does: Proposed a physics-based photometric bundle adjustment method for handling camera pose and 3D geometry estimation in non-Lambertian environments.
PhysORD: A Neuro-Symbolic Approach for Physics-infused Motion Prediction in Off-road Driving
Zhipeng Zhao, Chen Wang
Autonomous DrivingPhysics RelatedOrdinary Differential Equation
🎯 What it does: Proposes PhysORD, a neuro-symbolic method that embeds Euler-Lagrange equations into data-driven neural networks for motion prediction in off-road driving.