IROS 2024 Papers — Page 13
IEEE/RSJ International Conference on Intelligent Robots and Systems · 1581 papers
Renderable Street View Map-Based Localization: Leveraging 3D Gaussian Splatting for Street-Level Positioning
Howoong Jun, Songhwai Oh
Pose EstimationAutonomous DrivingGaussian SplattingImage
🎯 What it does: Using 3D Gaussian splatting to convert street-level 2D images into renderable 3D maps, and iteratively estimating the position and orientation of the query image by comparing the pose similarity between the rendered image and the query RGB image.
Repairing Neural Networks for Safety in Robotic Systems using Predictive Models
K. Majd, H. B. Amor
Safty and PrivacyRobotic Intelligence
🎯 What it does: Propose a safe-aware robot learning method based on a predictive model, first learning policies through behavioral cloning and then repairing the neural network to satisfy safety constraints.
REPeat: A Real2Sim2Real Approach for Pre-acquisition of Soft Food Items in Robot-assisted Feeding
N. Ha, T. Bhattacharjee
Robotic IntelligenceWorld ModelImage
🎯 What it does: Developed the REPeat framework, utilizing pre-capture actions (push, cut, flip) to enhance the success rate of biting soft foods in robot-assisted feeding;
Representing 3D sparse map points and lines for camera relocalization
Thuan Bui Bach, Joo-Ho Lee
Pose EstimationGraph Neural NetworkTransformerSimultaneous Localization and MappingGraph
🎯 What it does: Learn a lightweight neural network to simultaneously represent 3D point and line features, enhancing camera relocalization accuracy through various learning mappings.
Research of calibration method for fusion of LDS sensor and ToF low-cost sensor
Jiahui Zhu, Dongtai Liang
OptimizationPoint Cloud
🎯 What it does: Proposes an extrinsic calibration method for LDS sensors and ToF depth cameras using three cylinders, fitting the ellipse center points and cylinder axes with RANSAC, then solving the transformation matrix between the two sensors via Powell and BFGS optimization.
Research on Autonomous Navigation of Dual-mode Wheel-legged Robot
Wen Wang, Minzhou Luo
OptimizationRobotic Intelligence
🎯 What it does: Proposed and implemented a dual-mode navigation system based on a robot energy consumption model, first evaluating obstacle traversability and preprocessing 2D grid maps; integrated the energy consumption model into the A* evaluation function, improved the eight-connected expansion mode to search neighboring nodes based on obstacle characteristics; intelligently switched between crossing and detouring modes in obstacle crossing areas according to the minimum energy consumption principle, ultimately completing the dual-mode robot navigation system.
Resource-Aware Collaborative Monte Carlo Localization with Distribution Compression
Nicky Zimmerman, Jérôme Guzzi
Computational EfficiencySimultaneous Localization and Mapping
🎯 What it does: Propose a collaborative global localization method under computational and communication constraints, reducing information exchange and computational costs.
Response Improvement of Hydraulic Robotic Joints via a Force Servo and Inverted Pendulum Demo
Ryo Arai, K. Ono
Robotic IntelligencePhysics Related
🎯 What it does: Designed a force servo for a hydraulic robot joint and applied it to an inverted pendulum demonstration.
Retargeting Human Facial Expression to Human-like Robotic Face through Neural Network Surrogate-based Optimization
Bowen Wu, Hiroshi Ishiguro
OptimizationRobotic Intelligence
🎯 What it does: Developed a data-driven method to achieve zero-human intervention redirection of human facial expressions to robot faces.
Rethinking 3D Geometric Object Features for Enhancing Skeleton-based Action Recognition
Yuankai Wu, Eckehard G. Steinbach
RecognitionGraph Neural NetworkGraph
🎯 What it does: Propose a method to effectively integrate 3D geometric object features into skeletal data, implemented using graph convolutional networks (GCN).
Revisiting Reward Design and Evaluation for Robust Humanoid Standing and Walking
Bart Jaap van Marum, Alan Fern
Robotic IntelligenceReinforcement LearningBenchmark
🎯 What it does: Proposed a low-cost, quantifiable benchmark framework for evaluating and comparing the performance of standing and walking controllers in the real world, designed a minimal constraint reward function to train these controllers, and conducted experimental verification on the Digit robot.
Revolutionizing Battery Disassembly: The Design and Implementation of a Battery Disassembly Autonomous Mobile Manipulator Robot(BEAM-1)
Yanlong Peng, Ming Chen
Robotic IntelligenceLarge Language ModelMultimodality
🎯 What it does: Developed a neuro-symbolic AI-based autonomous mobile manipulator robot, BEAM-1, for high-precision disassembly of electric vehicle batteries.
Reward-Driven Automated Curriculum Learning for Interaction-Aware Self-Driving at Unsignalized Intersections
Zeng Peng, Jun Ma
Autonomous DrivingReinforcement Learning
🎯 What it does: Proposes a reward-driven automated curriculum reinforcement learning method for interactive perception-based autonomous driving control at unsignalized intersections, addressing uncertainty from surrounding vehicles;
Reward-field Guided Motion Planner for Navigation with Limited Sensing Range
J. Bayer, J. Faigl
Autonomous DrivingReinforcement Learning
🎯 What it does: Improve the navigation and exploration planning efficiency for ground vehicles in unknown or partially known environments by proposing a method based on reusable reward functions guiding a fast sampling-based motion planner.
Riemannian Flow Matching Policy for Robot Motion Learning
Max Braun, Tamim Asfour
Robotic IntelligenceDiffusion modelFlow-based ModelMultimodality
🎯 What it does: Proposes the Riemannian Flow Matching Strategy (RFMP) for learning and synthesizing robot audiovisual motor policies, demonstrating its applicability under state and visual conditions.
RISE: 3D Perception Makes Real-World Robot Imitation Simple and Effective
Chenxi Wang, Cewu Lu
Robotic IntelligenceTransformerDiffusion modelPoint Cloud
🎯 What it does: Developed an end-to-end baseline RISE that directly predicts continuous robotic actions using single-view point clouds, achieving more precise spatial perception and operational control.
Risk-Averse Planning and Plan Assessment for Marine Robots*
Mahya Mohammadi Kashani, Andrzej Wasowski
OptimizationRobotic Intelligence
🎯 What it does: Propose a method for marine robot task planning that first generates diverse high-level plans and then evaluates them in low-level simulations to select the optimal and most reliable candidate plans.
Risk-Aware Non-Myopic Motion Planner for Large-Scale Robotic Swarm Using CVaR Constraints
Xuru Yang, Chang Liu
OptimizationRobotic Intelligence
🎯 What it does: Proposed a non-greedy swarm motion planner (ROVER) based on conditional value-at-risk (CVaR) constraints, utilizing finite-time horizon model predictive control (FTMPC) and Gaussian Mixture Model (GMM) macro-state descriptions to ensure collision safety for large-scale robot swarms in complex environments.
RMap: Millimeter-Wave Radar Mapping Through Volumetric Upsampling
Ajay Narasimha Mopidevi, Christoffer Heckman
GenerationData SynthesisTransformerPoint Cloud
🎯 What it does: Using RMap to generate accurate 3D maps from sparse radar point clouds
RNR-Nav: A Real-World Visual Navigation System Using Renderable Neural Radiance Maps
Minsoo Kim, Songhwai Oh
Autonomous DrivingNeural Radiance FieldSimultaneous Localization and MappingImage
🎯 What it does: Proposed a visual localization and navigation framework called RNR-Nav for real-world environments, which directly integrates observed visual information into a bird's-eye view map and improves RNR-Map to RNR-Map++ to reduce information loss.
Road Boundary Estimation Using Sparse Automotive Radar Inputs
Aaron Kingery, Dezhen Song
Autonomous DrivingOptimization
🎯 What it does: This paper proposes a road boundary estimation method based on sparse automotive radar signals
Roadmaps with Gaps over Controllers: Achieving Efficiency in Planning under Dynamics
A. Sivaramakrishnan, Kostas E. Bekris
Autonomous DrivingOptimizationComputational EfficiencyRobotic Intelligence
🎯 What it does: Construct a 'Roadmap with Gaps' using a learning controller and combine it with a tree-based sampling planner to enhance the computational efficiency of motion planning for mobile robots in complex dynamic environments.
Roaming with Robots: Utilizing Artificial Curiosity in Global Path Planning for Autonomous Mobile Robots
N. Spielbauer, R. Dillmann
OptimizationRobotic IntelligenceGraph
🎯 What it does: Proposes a multi-objective genetic algorithm path planning method based on artificial curiosity, where the robot dynamically determines information-gathering paths during task intervals, aiming to maximize information gain and fully utilize available time.
RoboCop: A Robust Zero-Day Cyber-Physical Attack Detection Framework for Robots
Upinder Kaur, R. Voyles
Anomaly DetectionRobotic Intelligence
🎯 What it does: Proposed the RoboCop framework for detecting zero-day attacks in robotic systems
RoboGuardZ: A Scalable Zero-Shot Framework for Detecting Zero-Day Malware in Robots
Upinder Kaur, R. Voyles
Anomaly DetectionComputational EfficiencyRobotic Intelligence
🎯 What it does: Proposed the RoboGuardZ framework, using zero-shot learning to detect zero-day malware on robots.
Robot Active Vision-Based Path Planning for Localization Improvement in Indoor Environments
Sotirios Barlakas, Dimitrios Tzovaras
Robotic IntelligenceWorld Model
🎯 What it does: Proposed an active vision-based path planning method that generates efficient trajectories minimizing localization error and enhancing navigation performance through artificial potential fields (APF), KB-RRT, and weight-adaptive MPC.
Robot Design Optimization with Rotational and Prismatic Joints using Black-Box Multi-Objective Optimization
Kento Kawaharazuka, Masayuki Inaba
Optimization
🎯 What it does: Propose a robot design optimization method that combines rotational and translational joints, utilizing black-box multi-objective optimization to automatically generate robot structures with minimized number of joints and link lengths while satisfying task requirements.
Robot Generating Data for Learning Generalizable Visual Robotic Manipulation
Yunfei Li, Yi Wu
Data SynthesisRobotic IntelligenceReinforcement Learning
🎯 What it does: Proposed and implemented a robot self-teaching framework that utilizes the robot itself to generate effective training data and train visual control strategies;
Robot Guided Evacuation with Viewpoint Constraints
Gong Chen, Marcel Bartholomeus Prasetyo
OptimizationRobotic Intelligence
🎯 What it does: Proposed a viewpoint-based nonlinear model predictive control (MPC) algorithm to guide robots in tracking and leading cooperative human targets in emergency scenarios.
Robot Shape and Location Retention in Video Generation Using Diffusion Models
Peng Wang, Minh Huy Pham
GenerationRobotic IntelligenceConvolutional Neural NetworkDiffusion modelVideo
🎯 What it does: Proposed a Diffusion model specifically tailored for shape and position preservation in mobile robots, capable of generating videos that precisely retain the robot's morphology and positional information.
Robot Swarm Control Based on Smoothed Particle Hydrodynamics for Obstacle-Unaware Navigation
Michikuni Eguchi, Takefumi Hiraki
Robotic IntelligencePhysics Related
🎯 What it does: Propose a robot swarm control method based on Smoothed Particle Hydrodynamics (SPH), utilizing indirect obstacle detection to achieve navigation without obstacle perception;
Robot Synesthesia: A Sound and Emotion Guided Robot Painter
Vihaan Misra, Jean Oh
Robotic IntelligenceAuto EncoderTextMultimodalityAudio
🎯 What it does: Proposes the Robot Synesthesia method, which uses sound and speech to guide a robot in painting.
Robot Traversability Prediction: Towards Third-Person-View Extension of Walk2Map with Photometric and Physical Constraints
J. Liang, Kanji Tanaka
Robotic IntelligenceSimultaneous Localization and MappingImage
🎯 What it does: Propose a third-person perspective-based robot traversability prediction method called Walk2Map++, which addresses visual uncertainty issues by integrating photometric constraints (occlusion ordering) and physical constraints (collision avoidance);
Robot-Enabled Machine Learning-Based Diagnosis of Gastric Cancer Polyps Using Partial Surface Tactile Imaging
Siddhartha Kapuria, F. Alambeigi
ClassificationData SynthesisBiomedical Data
🎯 What it does: Proposed and evaluated a vision-based tactile sensor (VTS) and a machine learning algorithm that utilizes texture features for tumor classification, combined with seven-degree-of-freedom robotic manipulation to achieve automatic data acquisition and diagnosis of polymorphic gastric cancer tumors.
RobotGraffiti: An AR tool for semi-automated construction of workcell models to optimize robot deployment
Krzysztof Zielinski, M. Kjærgaard
OptimizationRobotic IntelligencePoint Cloud
🎯 What it does: Developed a semi-automatic tool based on augmented reality for building work cell models and optimizing robot deployment positions.
Robotic Measurement for Electrical Property of Polymers by Force-Sensing Robot toward Materials Lab-Automation
Yuki Asano, Junichiro Shiomi
Robotic IntelligencePhysics Related
🎯 What it does: An automated platform integrating a force-sensing robot with a measurement instrument control system was constructed, addressing technical challenges such as placement stability of polymer films during insertion into measurement instruments, cross-platform communication control, and the film transfer environment.
Robotic Object Insertion with a Soft Wrist through Sim-to-Real Privileged Training
Yuni Fuchioka, Masashi Hamaya
Domain AdaptationRobotic Intelligence
🎯 What it does: The study investigates using a soft wrist robot to perform insertion tasks involving contact-rich objects in unstructured environments, and proposes a sim-to-real method based on privileged training.
Robotic valve turning: axial misalignment estimation from reaction torques
Gautami Golani, Domenico Campolo
Robotic IntelligencePhysics Related
🎯 What it does: Proposed a simplified quasi-static model to predict the relationship between the base reaction torque and axial misalignment when a circular valve rotates in a two-point contact gripper.
ROBOVERINE: A human-inspired neural robotic process model of active visual search and scene grammar in naturalistic environments
R. Grieben, Gregor Schöner
Robotic Intelligence
🎯 What it does: Propose the ROBOVERINE model to achieve selective attention and scene grammar reasoning in robot active vision, addressing various challenges in visual attention and spatial analysis.
Robust Agility via Learned Zero Dynamics Policies
Noel Csomay-Shanklin, Aaron D. Ames
OptimizationRobotic Intelligence
🎯 What it does: Designed a robust and agile controller for hybrid underactuated systems and proposed the Zero Dynamics Policies method.
Robust and Safe Task-Driven Planning and Navigation for Heterogeneous Multi-Robot Teams with Uncertain Dynamics
Tianyang Pan, L. Kavraki
Robotic Intelligence
🎯 What it does: Propose a control framework to address motion planning for heterogeneous multi-robot teams with uncertain dynamics, incorporating decentralized feedback control and sampling-based motion planning.
Robust Backstepping Controller with Adaptive Sliding Mode Observer for a Tilt-Augmented Quadrotor With Uncertainty Using SO(3)
Sathyanarayanan Seshasayanan, S. R. Sahoo
Robotic Intelligence
🎯 What it does: Designed an adaptive super-twisting sliding mode observer to achieve finite-time estimation of uncertain terms, and constructed a backward tracking controller based on SO(3), enabling the tilted thrust quadrotor to achieve exponential convergence under disturbances and inertia variations.
Robust Imitation Learning for Mobile Manipulator Focusing on Task-Related Viewpoints and Regions
Yutaro Ishida, Hiroshi Bito
Robotic IntelligenceReinforcement LearningImage
🎯 What it does: Propose a robust imitation learning method that learns task-related perspectives and spatial regions under multi-view settings for mobile manipulators, enhancing the generalization ability of visual observations.
Robust Multi-Camera BEV Perception: An Image-Perceptive Approach to Counter Imprecise Camera Calibration
Rundong Sun, Yi Yang
Autonomous DrivingImagePoint Cloud
🎯 What it does: Proposed a robust multi-camera BEV perception network combining dual-space position encoding (DSPE) and image-awareness to improve tolerance to camera calibration errors and pose fluctuations.
Robust Online Epistemic Replanning of Multi-Robot Missions
Lauren Bramblett, N. Bezzo
Robotic Intelligence
🎯 What it does: Proposes a two-stage framework, first using a centralized planner to allocate multi-robot tasks through intermittent meetings, then employing decentralized cognitive planning and Monte Carlo Tree Search (MCTS) for adaptive replanning;
Robust Partitioned Visual Servoing for Aerial Manipulation Utilizing Controllable-space Image Planning and Adaptive Image Representation
M. Soltanshah, Kamal Gupta
OptimizationRobotic IntelligenceImage
🎯 What it does: A robust visual servoing method for image space planning within a controllable space is proposed, addressing the infeasibility of piecewise linear camera trajectories on drone operation platforms caused by underactuation and auxiliary tasks.
Robust Precision Landing of a Quadrotor with Online Temporal Scaling Adaptation of Dynamic Movement Primitives
Kongkiat Rothomphiwat, P. Manoonpong
Robotic Intelligence
🎯 What it does: An adaptive trajectory planning method based on online temporal scaling dynamic movement primitives (DMP) is proposed to achieve robust and precise landing of quadrotor drones on stationary and moving ground targets in the presence of disturbances.
Robust Two-View Geometry Estimation with Implicit Differentiation
Vladislav A. Pyatov, Stamatios Lefkimmiatis
Pose Estimation
🎯 What it does: Proposes a two-view geometry estimation framework based on differentiable robust loss function fitting, treating the robust fundamental matrix estimation as an implicit layer, and constructing an end-to-end trainable unified pipeline
Robust-Adaptive Two-Loop Control for Robots with Mixed Rigid-Elastic Joints
T. Hua, Filippo Sanfilippo
Robotic Intelligence
🎯 What it does: Proposed a robust adaptive dual-loop control algorithm for hybrid rigid-flexible joint robots
Robustifying Model-Based Locomotion by Zero-order Stochastic Nonlinear Model Predictive Control with Guard Saltation Matrix
Sotaro Katayama, Masaya Kinoshita
Robotic Intelligence
🎯 What it does: Propose a stochastic/robust nonlinear model predictive control (NMPC) based on a zeroth-order algorithm, utilizing a gate salting matrix and extended Kalman filter covariance update to enhance the robustness of model-based multi-legged locomotion against contact uncertainty.
Robustness Study of Optimal Geometries for Cooperative Multi-Robot Localization
Mathilde Theunissen, Philippe Martinet
OptimizationRobotic Intelligence
🎯 What it does: This paper conducts a systematic study on the robustness of different formations in multi-robot collaborative localization, deriving the necessary and sufficient conditions for formation robustness in 2D space, and verifying through simulations the robust advantages of regular polygon formations and the defects of square formations.
Rocket Landing Control with Random Annealing Jump Start Reinforcement Learning
Yuxuan Jiang, Changwu Zhang
Autonomous DrivingReinforcement Learning
🎯 What it does: Using reinforcement learning (RL) via the Random Annealing Jump Start (RAJS) method to control the real-time landing of a nonlinear underactuated rocket.
ROG-Map: An Efficient Robocentric Occupancy Grid Map for Large-scene and High-resolution LiDAR-based Motion Planning
Yunfan Ren, Fu Zhang
Autonomous DrivingComputational EfficiencySimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposed ROG-Map, a unified grid-based occupancy grid map that maintains a local map as the robot moves and achieves incremental obstacle dilation, improving the efficiency of high-resolution LiDAR map updates in large-scale environments.
Roofus: Learning-based Robotic Moisture Mapping on Flat Rooftops with Ground Penetrating Radar
Kevin Lee, Chen Feng
Robotic IntelligenceTransformerSupervised Fine-TuningSimultaneous Localization and Mapping
🎯 What it does: Developed an integrated robotic humidity detection system for flat roofs, combining remote-controlled robots, ground-penetrating radar (GPR) data, deep learning processing, and automated map generation.
ROS-lite2: Autonomous-driving Software Platform for Clustered Many-core Processor
Yuta Tajima, Takuya Azumi
Autonomous Driving
🎯 What it does: This paper proposes the ROS-lite2 software platform based on the ROS 2 framework and multi-core processors, aiming to enhance the flexibility and ease of deployment for autonomous vehicles, while supporting complex function integration and intuitive operation, and reducing dependency on hardware knowledge.
Rotograb: Combining Biomimetic Hands with Industrial Grippers using a Rotating Thumb
Arnaud Bersier, Robert K. Katzschmann
Robotic IntelligenceReinforcement LearningImage
🎯 What it does: Developed a flexible grasping system called Rotograb, which combines human hands with industrial grippers through a rotating thumb, and integrates remote control operation with autonomous grasping based on reinforcement learning.
RPMArt: Towards Robust Perception and Manipulation for Articulated Objects
Junbo Wang, Cewu Lu
Pose EstimationDomain AdaptationRobotic IntelligencePoint Cloud
🎯 What it does: Propose the RPMArt framework, which estimates joint parameters from noisy point clouds and generates robust actions to achieve perception and manipulation of articulated objects.
RT-Grasp: Reasoning Tuning Robotic Grasping via Multi-modal Large Language Model
Jinxuan Xu, Liangjun Zhang
Pose EstimationRobotic IntelligenceTransformerLarge Language ModelVision-Language-Action ModelMultimodalityChain-of-Thought
🎯 What it does: The study applies the reasoning capabilities of large language models (LLM) to robot grasping tasks by incorporating a reasoning phase during training to generate numerical predictions.
RT-RRT: Reverse Tree Guided Real-Time Path Planning/Replanning in Unpredictable Dynamic Environments
Bo Cui, Shi Zhang
Autonomous DrivingOptimization
🎯 What it does: Proposed a reverse tree-guided real-time path planning/replanning algorithm (RT-RRT) that can efficiently perform navigation tasks in unpredictable dynamic environments.
RTTF: Rapid Tactile Transfer Framework for Contact-Rich Manipulation Tasks
Qiwei Wu, Yunjiang Lou
Domain AdaptationRobotic IntelligenceAuto EncoderImage
🎯 What it does: Propose a fast tactile transfer framework (RTTF) that achieves sim-to-real transfer for optical tactile images and enables robust tactile servo control under limited paired data.
S-BUN: Soft Bifunctional Utility Module for Robot Sensing and Signaling
Suksakaow Mahuttanatan, P. Manoonpong
Robotic Intelligence
🎯 What it does: Proposed a soft dual-functional module S-BUN that integrates non-contact distance perception, touch perception, and state signal transmission functions for robots.
Safe and Efficient Auto-tuning to Cross Sim-to-real Gap for Bipedal Robot
Yidong Du, Qiang Huang
Domain AdaptationOptimizationHyperparameter SearchData-Centric LearningRobotic Intelligence
🎯 What it does: Train a parameter search model using automatic tuning technology, adjust simulation parameters with real data, further formulate the problem as a discrete distribution issue, and incorporate a dataset to validate the model, achieving safe and efficient simulation parameter tuning for BITeno.
Safe CoR: A Dual-Expert Approach to Integrating Imitation Learning and Safe Reinforcement Learning Using Constraint Rewards
Hyeokjin Kwon, Songhwai Oh
Autonomous DrivingReinforcement LearningMixture of Experts
🎯 What it does: Propose a Safe CoR framework that combines reward expert and safety expert demonstrations to simultaneously optimize performance and safety constraints.
Safe Imitation Learning of Nonlinear Model Predictive Control for Flexible Robots
Shamil Mamedov, J. Swevers
OptimizationComputational EfficiencyRobotic Intelligence
🎯 What it does: Proposes a safe approximate nonlinear model predictive control framework based on imitation learning and predictive safety filters to achieve fast control of flexible robots.
Safe multi-agent reinforcement learning for bimanual dexterous manipulation
Weishu Zhan, Peter Chin
Safty and PrivacyRobotic IntelligenceReinforcement Learning
🎯 What it does: Proposed a multi-agent reinforcement learning algorithm called MAC-PAO for safe bimanual coordination, and conducted experiments on various tasks with safety constraints
Safe Offline-to-Online Multi-Agent Decision Transformer: A Safety Conscious Sequence Modeling Approach
A. Shah, Xin Fu
TransformerReinforcement LearningSequentialBenchmark
🎯 What it does: Proposed a safety-aware offline-to-online multi-agent decision Transformer framework SO2-MADT, achieving safety-priority multi-agent reinforcement learning.
Safe Reinforcement Learning via Hierarchical Adaptive Chance-Constraint Safeguards
Zhaorun Chen, Chengliang Liu
Reinforcement Learning
🎯 What it does: Proposed an adaptive chance-constrained safety assurance algorithm (ACS), which uses safety recovery rate as a chance constraint to iteratively ensure safety during exploration and after convergence.
Safety-critical Autonomous Inspection of Distillation Columns using Quadrupedal Robots Equipped with Roller Arms
Jaemin Lee, A. Ames
Robotic Intelligence
🎯 What it does: Proposes a full-process framework based on quadruped robots and roller arms, achieving autonomous inspection of multi-level distillation tower trays, integrating key motion components such as gait, safe dynamic transfer between trays, and intermediate motion.
Safety-First Tracker: A Trajectory Planning Framework for Omnidirectional Robot Tracking
Yue Lin, Huchuan Lu
OptimizationRobotic Intelligence
🎯 What it does: Proposes a safety-priority trajectory planning framework (SF-Tracker) for omnidirectional autonomous tracking robots, which separates robot position and orientation for step-by-step planning. First, it constructs a reference path that is independent of conflicts and occlusions, then performs safety trajectory optimization, and designs an orientation planner to ensure target visibility.
SAID-NeRF: Segmentation-AIDed NeRF for Depth Completion of Transparent Objects
Avinash Ummadisingu, Kuniyuki Takahashi
SegmentationDepth EstimationNeural Radiance Field
🎯 What it does: Using a vision foundation model (VFM) for zero-shot, unlabeled segmentation to guide NeRF reconstruction, and enhancing the robustness of depth completion for transparent objects through simultaneous reconstruction and expansion of semantic fields.
Sampling-based Motion Planning for Optimal Probability of Collision under Environment Uncertainty
Hao Lu, Rahul Shome
OptimizationRobotic Intelligence
🎯 What it does: The study addresses motion planning under environmental uncertainty, proposing a theoretically sound optimal collision probability model, and achieving path planning through a hybrid method combining mixed-integer linear programming (MILP) with greedy search.
Satellite-Model-Free Deep Learning based Pose Estimation of Non-cooperative Satellite and Tracking using Navigation Filter
Shubham Shukla, Titas Bera
Pose EstimationConvolutional Neural NetworkImage
🎯 What it does: Proposed a direct pose estimation architecture based on a monocular camera without a 3D model, and implemented an adaptive navigation filter on this foundation to achieve continuous pose tracking.
Saturation in the Null-Space (SNS) for Tele-operated Surgery: Prioritized Motion Control for RCM and Joint Limit Constraints
Sreekanth Kana, Shashank Sharma
Robotic Intelligence
🎯 What it does: Applying the SNS algorithm in minimally invasive surgical robots to achieve task prioritization and coordination, focusing on ensuring RCM constraints while considering joint limits
SAVOR: Sonar-Aided Visual Odometry and Reconstruction for Autonomous Underwater Vehicles*
J. Coffelt, Bilal Wehbe
Robotic IntelligenceSimultaneous Localization and MappingImageMultimodality
🎯 What it does: Propose a hybrid visual odometry and reconstruction method using only a camera and multibeam sonar, suitable for autonomous underwater vehicles (AUVs) that do not rely on IMU/DVL.
Scalability of Platoon-based Coordination for Mixed Autonomy Intersections
Zhongxia Yan, Cathy Wu
Autonomous DrivingReinforcement Learning
🎯 What it does: Investigated the scalability of platoons' cooperative control in mixed automated intersections, and compared two control methods based on MPC and RL;
Scalable Multi-Agent Reinforcement Learning for Warehouse Logistics with Robotic and Human Co-Workers
Aleksandar Krnjaic, Stefano V. Albrecht
OptimizationReinforcement LearningAgentic AI
🎯 What it does: Proposes a hierarchical multi-agent reinforcement learning framework to coordinate robots and human pickers in a warehouse to maximize picking efficiency.
Scalable Network and Adaptive Refinement Module for 6D Pose Estimation of Diverse Industrial Components*
Kun Qian, Xianwen Kong
Pose EstimationConvolutional Neural NetworkImage
🎯 What it does: Proposes a scalable network architecture based on EfficientNet, combined with an enhanced feature pyramid network and an adaptive pose refinement module for 6D pose estimation of industrial components.
Scalable Networked Feature Selection with Randomized Algorithm for Robot Navigation
Vivek Pandey, N. Motee
OptimizationRobotic Intelligence
🎯 What it does: For robot teams navigating in unknown environments, the study focuses on algorithms that sparsely select visual features for localization, using simulations of robot trajectories during a prediction period to evaluate feature importance.
Scale Disparity of Instances in Interactive Point Cloud Segmentation
Chenrui Han, Yue Wang
SegmentationTransformerPoint Cloud
🎯 What it does: Proposed an interactive point cloud segmentation model called ClickFormer, capable of precisely segmenting instances of thing and stuff categories.
SCANet: Correcting LEGO Assembly Errors with Self-Correct Assembly Network
Yuxuan Wan, Hao Dong
Anomaly DetectionContrastive LearningImage
🎯 What it does: Proposes a self-correcting assembly network, SCANet, for the task of identifying and correcting LEGO assembly errors
SceneSense: Diffusion Models for 3D Occupancy Synthesis from Partial Observation
Alec Reed, Christoffer Heckman
GenerationData SynthesisDiffusion modelImage
🎯 What it does: Proposed a real-time 3D diffusion model called SceneSense, which can synthesize 3D occupancy information from local observations using only a single RGB-D camera and an ongoing occupancy map, predicting occluded or out-of-view geometry for future planning and control.
Scheduling of Robotic Cellular Manufacturing Systems with Timed Petri Nets and Reinforcement Learning
ZhuTao Yao, Shaohua Yu
OptimizationRobotic IntelligenceReinforcement Learning
🎯 What it does: Proposed a Q-learning scheduling method based on Petri nets for efficient scheduling in robot cell manufacturing systems.
SCOML: Trajectory Planning Based on Self-Correcting Meta-Reinforcement Learning in Hybrid Terrain for Mobile Robot
Andong Yang, Yu Hu
Robotic IntelligenceMeta LearningReinforcement Learning
🎯 What it does: Propose a trajectory planning network capable of handling mixed terrains, and introduce a self-correcting structure based on historical planning data to enhance safety; employ a two-phase offline meta-reinforcement learning training scheme to enable the network to learn from pre-collected suboptimal data, reducing the occurrence of dangerous planning.
SCP: Soft Conditional Prompt Learning for Aerial Video Action Recognition
Xijun Wang, Dinesh Manocha
RecognitionPrompt EngineeringMixture of ExpertsVideo
🎯 What it does: Develop the Soft Conditional Prompt Learning (SCP) method for aerial video action recognition, predicting actions for each subject and enhancing the model's focus on action descriptions through prompt learning.
SculptDiff: Learning Robotic Clay Sculpting from Humans with Goal Conditioned Diffusion Policy
Alison Bartsch, A. Farimani
Robotic IntelligenceDiffusion modelPoint Cloud
🎯 What it does: Proposed the SculptDiff framework, which uses point cloud state observations to achieve goal-conditioned sculpting control of 3D deformable objects such as clay.
SD-Net: Symmetric-Aware Keypoint Prediction and Domain Adaptation for 6D Pose Estimation In Bin-picking Scenarios
Ding-Tao Huang, Long Zeng
Pose EstimationDomain AdaptationPoint Cloud
🎯 What it does: Propose a new 6D pose estimation network SD-Net, combining symmetric-aware keypoint prediction with self-training domain adaptation technology.
SDFT: Structural Discrete Fourier Transform for Place Recognition and Traversability Analysis
Ayumi Umemura, Kazuya Yoshida
RecognitionRobotic Intelligence
🎯 What it does: Proposes a method based on structural discrete Fourier transform (SDFT) for scene recognition and traversability analysis of ground robots in irregular environments.
SDGE: Stereo Guided Depth Estimation for 360°Camera Sets
Jialei Xu, Xianming Liu
Depth EstimationSupervised Fine-TuningImageVideo
🎯 What it does: Propose a stereo-guided depth estimation method that leverages multi-view stereo results and self-calibrated camera poses to enhance the quality of full-image depth estimation for 360° camera sets.
SDPL-SLAM: Introducing Lines in Dynamic Visual SLAM and Multi-Object Tracking
Argyris Manetas, P. Maragos
Object TrackingSimultaneous Localization and MappingOptical FlowImageVideo
🎯 What it does: We propose a new dynamic visual SLAM system that can simultaneously estimate the camera pose and rigid object motion, fully utilizing point features and line segment features in all algorithm stages.
SDTrack:Spatially decoupled tracker for visual tracking
Zihao Xia, Zhiquan Wang
Object TrackingTransformer
🎯 What it does: Proposes a spatially decoupled visual tracker called SDTrack, which includes a query selection module, an incremental addition of Box-to-Pixel relative position information in cross-attention, and an alignment loss to enhance the matching effectiveness between classification and localization.
SE(3) Linear Parameter Varying Dynamical Systems for Globally Asymptotically Stable End-Effector Control
Sunan Sun, Nadia Figueroa
OptimizationRobotic IntelligenceOrdinary Differential Equation
🎯 What it does: Proposed Quaternion-DS, extending the LPV-DS framework to quaternion space to learn DS motion strategies for attitude, and combined it with position control to achieve the complete SE(3) LPV-DS;
Search-based Strategy for Spatio-Temporal Environmental Property Restoration
Amel Nestor Docena, Alberto Quattrini Li
OptimizationRobotic Intelligence
🎯 What it does: Proposes a search-based strategy for robots with limited battery life to continuously plan visits to regions of interest and charging in known environments, minimizing the time during which environmental attributes (e.g., air quality) fall below a threshold.
SeeBelow: Sub-dermal 3D Reconstruction of Tumors with Surgical Robotic Palpation and Tactile Exploration
Raghava Uppuluri, Yu She
Robotic IntelligenceImageBiomedical Data
🎯 What it does: Three-dimensional surface reconstruction of subcutaneous tumors in multi-layered tissues using a robotic palpation probe and visual-guided tactile navigation strategy.
Seg2Grasp: A Robust Modular Suction Grasping in Bin Picking
Hye Jung Yoon, Byoung-Tak Zhang
Robotic IntelligenceTransformerSupervised Fine-TuningVision Language Model
🎯 What it does: Proposes Seg2Grasp, a modular pipeline for achieving robust suction grasping in dynamic cluttered bin environments.
Segmented Safety Docking Control for Mobile Self-Reconfigurable Robots
Zhi Zheng, Jianchuan Ye
Robotic IntelligenceSimultaneous Localization and Mapping
🎯 What it does: Propose a segmented safe docking control framework based on global localization and local perception, achieving stable and reliable reconfiguration of mobile self-reconfigurable robots (MSRR) in real-world environments.
Self Supervised Detection of Incorrect Human Demonstrations: A Path Toward Safe Imitation Learning by Robots in the Wild
Noushad Sojib, M. Begum
Safty and PrivacyRobotic IntelligenceSequentialBenchmark
🎯 What it does: Created the Layman V1.0 dataset and proposed the Behavior Cloning for Error Detection (BED) framework to automatically detect and discard erroneous demonstrations, thereby enhancing the safety of imitation learning.
Self-Assessment of Robotic Laboratory and Equipment Readiness Using Large Language Models and Robotic Data Capture
Stefan Ilić, Josie Hughes
Robotic IntelligenceTransformerLarge Language ModelImageAudio
🎯 What it does: Integrate large language models with robot data capture to explore automated robot laboratory readiness assessment methods.
Self-reconfiguration Strategies for Space-distributed Spacecraft
Tianle Liu, Panfeng Huang
Robotic IntelligenceReinforcement Learning
🎯 What it does: Designed a distributed on-orbit spacecraft assembly algorithm, and learned the module processing sequence through a framework combining imitation learning and reinforcement learning, subsequently achieving self-reconfiguration tasks via robotic arm motion planning.
Self-Selecting Semi-Supervised Transformer-Attention Convolutional Network for Four Class EEG-Based Motor Imagery Decoding
Han Wei Ng, Cuntai Guan
ClassificationConvolutional Neural NetworkTransformerAuto EncoderBiomedical Data
🎯 What it does: Propose a multi-class EEG motor imagery classification method based on Variational Autoencoder and Transformer-based Attention Convolutional Network (SSTACNet).
Self-supervised Monocular Depth Estimation in Challenging Environments Based on Illumination Compensation PoseNet
Shengyuan Hou, Mengyin Fu
Depth EstimationAutonomous DrivingTransformerVideo
🎯 What it does: Proposes a self-supervised monocular depth estimation unified framework capable of performing depth estimation under complex lighting conditions such as nighttime, rain, and snow.