ICRA 2024 Papers — Page 5
IEEE International Conference on Robotics and Automation · 1760 papers
Decentralized Lifelong Path Planning for Multiple Ackerman Car-Like Robots
Teng Guo, Jingjin Yu
Autonomous DrivingOptimization
🎯 What it does: Proposed a multi-robot lifelong path planning method for multi-lane non-holonomic robots in continuous domains, and implemented centralized and distributed planners that enable robots to precisely reach SE(2) target poses.
Decentralized Multi-Agent Active Search and Tracking when Targets Outnumber Agents
Arundhati Banerjee, Jeff Schneider
Object TrackingReinforcement Learning
🎯 What it does: Proposed a distributed multi-agent active search and tracking algorithm called DecSTER, applicable to scenarios where the number of targets exceeds the number of agents and communication is asynchronous.
Decentralized Multi-Agent Trajectory Planning in Dynamic Environments with Spatiotemporal Occupancy Grid Maps
Siyuan Wu, Javier Alonso-Mora
Autonomous DrivingOptimization
🎯 What it does: A decentralized trajectory planning framework is proposed for multiple micro-drones to achieve collision avoidance in environments containing static and dynamic obstacles.
Decentralized multi-phase formation control for cattle herding
Dac Dang Khoa Nguyen, A. Alempijevic
Robotic IntelligenceAgriculture Related
🎯 What it does: Propose a decentralized multi-stage control strategy for herding cattle using a group of robots, divided into an encircling phase and a driving phase;
Decentralized Multi-Robot Navigation for Autonomous Surface Vehicles with Distributional Reinforcement Learning
Xi Lin, Brendan Englot
Autonomous DrivingRobotic IntelligenceReinforcement Learning
🎯 What it does: Proposes a decentralized multi-robot collision avoidance strategy based on distributed reinforcement learning, considering interactions between multiple ASVs, static obstacles, and water currents.
Decision Making for Human-in-the-loop Robotic Agents via Uncertainty-Aware Reinforcement Learning
S. Singi, M. Ciocarlie
Robotic IntelligenceReinforcement Learning
🎯 What it does: Propose a decision-making method for semi-autonomous robots to request expert assistance based on reinforcement learning;
Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulation
Annie Xie, Chelsea Finn
Robotic IntelligenceBenchmark
🎯 What it does: Studied the generalization difficulty of imitation learning strategies under different environmental factors on simulated and real robots, and designed a simulation benchmark consisting of 19 tasks and 11 varying factors for controlled evaluation.
Deep Compliant Control for Legged Robots
A. Hartmann, Stelian Coros
Robotic IntelligenceReinforcement Learning
🎯 What it does: Introducing an explicit recovery phase during deep reinforcement learning training, enabling robots to gradually recover and achieve more natural and compliant balance restoration when encountering disturbances.
Deep Evidential Uncertainty Estimation for Semantic Segmentation under Out-Of-Distribution Obstacles
Siddharth Ancha, Nicholas Roy
SegmentationConvolutional Neural NetworkFlow-based ModelImage
🎯 What it does: Proposes a method for accurately estimating pixel-level uncertainty in semantic segmentation without real or synthetic OOD samples, leveraging shared pixel latent feature representations. A classification network predicts semantic label class distributions, normalizing flows estimate the probability density of features under the training distribution, and a Dirichlet-type evidence uncertainty framework simultaneously computes epistemic and aleatoric uncertainties in a single forward pass.
Deep Learning based acoustic measurement approach for robotic applications on orthopedics
Bangyu Lan, Kenan Niu
Robotic IntelligenceConvolutional Neural NetworkBiomedical DataUltrasound
🎯 What it does: Propose a deep learning-based A-mode ultrasound bone tracking method, utilizing CasAtt-UNet to automatically and robustly predict bone positions from 1D raw ultrasound signals during TKA surgery.
Deep Model Predictive Optimization
Jacob Sacks, Byron Boots
OptimizationReinforcement Learning
🎯 What it does: Proposed and evaluated a Deep Model Predictive Optimization (DMPO) framework, which learns the internal loop of MPC through experience-based learning, and compared it with traditional MPC and MFRL in agile trajectory tracking tasks for a real quadrotor.
DefFusion: Deformable Multimodal Representation Fusion for 3D Semantic Segmentation
Rongtao Xu, Xiaopeng Zhang
SegmentationAutonomous DrivingTransformerImageMultimodalityPoint CloudBenchmark
🎯 What it does: Proposes the DefFusion framework to achieve camera and LiDAR multi-modal feature fusion based on deformable Transformer, and designs a dynamic representation enhancement module to remove noise.
DefGoalNet: Contextual Goal Learning from Demonstrations for Deformable Object Manipulation
Bao Thach, A. Kuntz
Robotic Intelligence
🎯 What it does: Achieve shape servoing by learning the target shape of deformable objects through a few human demonstrations;
DeFlow: Decoder of Scene Flow Network in Autonomous Driving
Qingwen Zhang (KTH Royal Institute Technology), P. Jensfelt (KTH Royal Institute Technology)
Autonomous DrivingRecurrent Neural NetworkOptical FlowPoint Cloud
🎯 What it does: Proposed the DeFlow network for scene flow estimation on large-scale point clouds, refining the conversion from voxel features to point features using GRU
DeformNet: Latent Space Modeling and Dynamics Prediction for Deformable Object Manipulation
Chenchang Li, Huazhe Xu
Robotic IntelligenceNeural Radiance FieldWorld ModelPoint Cloud
🎯 What it does: Proposes a framework called DeformNet for manipulating deformable objects with infinite degrees of freedom, combining latent space modeling with temporal dynamic prediction;
Degenerate Motions of Multisensor Fusion-based Navigation
W. Lee, Guoquan Huang
Autonomous DrivingOptimizationSimultaneous Localization and Mapping
🎯 What it does: Analyzes the observability of a multi-sensor fusion navigation system, identifies 9 degenerate motions, and verifies their impact through numerical simulations.
Degradation Resilient LiDAR-Radar-Inertial Odometry
Morten C. Nissov, Kostas Alexis
Pose EstimationAutonomous DrivingSimultaneous Localization and MappingMultimodalityPoint Cloud
🎯 What it does: Proposed a tight-coupled LiDAR-Radar-Inertial fusion pose estimation method to mitigate the effects of LiDAR distortion;
DenseTact-Mini: An Optical Tactile Sensor for Grasping Multi-Scale Objects From Flat Surfaces
Won Kyung Do, Monroe Kennedy
Robotic Intelligence
🎯 What it does: Proposed a mini optical tactile sensor named DenseTact-Mini with a soft, smooth gel surface and synthetic fingernails, and designed three grasping strategies: adhesion-based tapping grasp, fingernail grasp utilizing rolling/sliding contact of the fingernail, and dual soft fingertip grasp;
DerainNeRF: 3D Scene Estimation with Adhesive Waterdrop Removal
Yunhao Li, Peidong Liu
RestorationNeural Radiance FieldImage
🎯 What it does: Remove water droplets from multi-view images by using an attention network to predict water droplet locations and training NeRF to reconstruct a clear 3D scene.
Design & Systematic Evaluation of Power Transmission Efficiency of an Ankle Exoskeleton for Walking Post-Stroke
Myles Cooper, C. J. Walsh
Robotic IntelligenceBiomedical Data
🎯 What it does: Designed a portable rigid ankle exoskeleton and tested it on stroke rehabilitation patients to validate its potential for future community use; subsequently proposed a systematic method to quantify power transmission loss at each stage from the battery to the user.
Design and Analysis of Soft Hybrid-Driven Manipulator with Variable Stiffness and Multiple Motion Patterns
Xin Fu, Xingang Zhao
Robotic Intelligence
🎯 What it does: Propose and implement a soft hybrid-driven robotic arm capable of continuous stiffness control and achieving omnidirectional bending and extension motion modes. By combining soft airbag actuation with tensile tendons, the system enables independent control of three-dimensional spatial position and stiffness. Experimental results demonstrate an elongation rate of 198%, a bending angle of 240°, and trajectory tracking experiments validate the accuracy of the kinematic model.
Design and Central Pattern Generator Control of a New Transformable Wheel-Legged Robot
Tyler Bishop, Konstantinos Karydis
Robotic Intelligence
🎯 What it does: Developed a deformable wheel-leg robot and designed a motion controller based on Central Pattern Generators (CPG) to enable navigation across various terrains.
Design and Characterization of a Soft Flat Tube Twisting Actuator
Hao Liu, Yonghua Chen
Robotic IntelligencePhysics Related
🎯 What it does: Proposed and tested a soft torsional actuator SFTTA composed of folded flat tubes and silicone lamination.
Design and evaluation of a modular robotic system for microsurgery
Jenireth Torrealba Molina, Etienne Burdet
Robotic Intelligence
🎯 What it does: Designed and evaluated a modular minimally invasive surgical robot system operated using standard surgical instruments
Design and Evaluation of a Reconfigurable 7-DOF Upper Limb Rehabilitation Exoskeleton with Gravity Compensation
Linliang Zheng, Qiang Zhang
OptimizationRobotic Intelligence
🎯 What it does: Designed and evaluated a reconfigurable 7-degree-of-freedom upper limb rehabilitation exoskeleton, which achieves left-right arm switching through reconfiguration and features gravity compensation capabilities.
Design and Evaluation of Motion Planners for Quadrotors in Environments with Varying Complexities
Y. Shao, Vijay Kumar
Robotic Intelligence
🎯 What it does: Studied the design and evaluation of a two-stage motion planner for quadrotors in various complexity environments
Design and Experimental Characterisation of a Novel Quasi-Direct Drive Actuator for Highly Dynamic Robotic Applications
C. A. Pérez-Díaz, Miguel López
Robotic Intelligence
🎯 What it does: Designed and experimentally verified a novel quasi-direct drive (QDD) actuator PULSE115-60, focusing on its mechanical design, dynamic performance parameters, and methods for measuring speed/torque bandwidth.
Design and Fabrication of a Novel Miniature Magnetic Gripper
Mengde Li, Miao Li
Robotic Intelligence
🎯 What it does: Designed and manufactured a four-fingered micro-magnetic gripper, and developed a data-driven kinematic model based on external magnetic field data for the opening and closing angle. The model's accuracy and the gripper's magnetic sensitivity were subsequently validated through experiments.
Design and Fabrication of String-driven Origami Robots
Peiwen Yang, Shuguang Li
Robotic Intelligence
🎯 What it does: Proposed a fast design and manufacturing method for rope-driven origami structures and robots, achieving the construction and testing of origami crawling robots and mechanical arms.
Design and Implementation of a Robotic Testbench for Analyzing Pincer Grip Execution in Human Specimen Hands
Nikolas J. Wilhelm, Rainer H. H. Burgkart
Robotic IntelligenceBiomedical Data
🎯 What it does: Designed and implemented a robotic testbed using eight force-controlled motors driving tendons, an optical tracking system, and torque sensors to demonstrate and measure the thumb pincer grip on a cadaver hand, as well as the relationship between tendon forces and grip strength.
Design and Implementation of A Robotized Hand-held Dissector for Endoscopic Pulmonary Endarterectomy
Runfeng Zhu, Qingxiang Zhao
Robotic Intelligence
🎯 What it does: Designed and implemented a handheld robotic steerable dissector for pulmonary artery endarterectomy, capable of gently removing deep layers of the pulmonary artery intima and providing aspiration and visualization functions.
Design and Modeling of A Compact Serial Variable Stiffness Actuator (SVSA-III) with Linear Stiffness Profile
Shuowen Yi, Zhao Guo
Robotic Intelligence
🎯 What it does: Designed and modeled a compact serial variable stiffness actuator (SVSA-III), achieving adjustable stiffness through a symmetric variable link mechanism and an Archimedean spiral displacement mechanism, capable of customizing a continuous linear stiffness curve.
Design and Modeling of a Nested Bi-cavity-based Soft Growing Robot for Grasping in Constrained Environments
Haochen Yong, Zhigang Wu
Robotic Intelligence
🎯 What it does: Proposed a nested dual-chamber structure soft growth robot (BIBOT) capable of navigation and grasping in confined environments
Design and Testing of a Multi-Module, Tetherless, Soft Robotic Eel
Robin Hall, C. Onal
Robotic Intelligence
🎯 What it does: Designed, fabricated, and evaluated a three-module, cable-free, flexible soft-bodied eel;
Design And Validation of a Variable Stiffness Spiral Cam Actuator (VS-SCA)
Matthew R. Auer, Hyunglae Lee
Robotic Intelligence
🎯 What it does: Designed and verified a multi-blade variable stiffness actuator (VS-SCA) for gait assistance in individuals with weakened ankle function.
Design and validation of slender extensible continuum robot for solar wing re-unfolding in aerospace
Pengyuan Wang, Jianwen Zhao
Robotic IntelligencePhysics Related
🎯 What it does: Designed and verified a deployable continuum robot for the re-deployment of solar arrays in space
Design and Visual Servoing Control of a Hybrid Dual-Segment Flexible Neurosurgical Robot for Intraventricular Biopsy
Jian Chen, Hongbin Liu
Robotic IntelligenceBiomedical Data
🎯 What it does: Proposed the dual-segment flexible robotic endoscope MicroNeuro for deep brain biopsy, enhancing motion accuracy and stability through image-based visual servoing, online Jacobian estimation, and constrained model predictive control.
Design Octree-Based Method to Improve Model-Mediated Teleoperation in Tactile Internet
Mads Antonsen, Qi Zhang
Robotic IntelligencePoint Cloud
🎯 What it does: Proposed a model-mediated teleoperation (MMT) system based on an octree model for spatial mapping of environmental impedance in tactile networks, avoiding continuous transmission of impedance and enhancing force feedback accuracy;
Design of a Front-enveloping Powered Exoskeleton Considering Optimal Distribution of Actuating Torques and Center of Mass
Jeongsu Park, Kyoungchul Kong
OptimizationRobotic Intelligence
🎯 What it does: Designed and verified a front-hugging powered exoskeleton, and determined the optimal front-back center of gravity position through optimization analysis.
Design of a Knee-joint Exoskeleton to Reduce Misalignment in Both the Sagittal and Coronal Planes
Shubhranil Sengupta, Jee-Hwan Ryu
Robotic Intelligence
🎯 What it does: Designed a knee exoskeleton capable of simultaneously reducing knee misalignment in the sagittal and coronal planes.
Design of A Rigid-soft Hybrid Robotic Glove with Force Sensing Function
Hexin Li, Kehan Ding
Robotic Intelligence
🎯 What it does: Designed a hard-soft hybrid robotic glove integrating rigid and flexible structures with force sensing capabilities.
Design of a Towing System by Multi Autonomous Sailboats*
Cheng Liang, Huihuan Qian
OptimizationRobotic Intelligence
🎯 What it does: Proposes a towing system design for multiple autonomous sailboats and conducts experiments to find optimal sail and rudder control strategies to enhance towing force and tangential success rate.
Design of Embodied Mediator Haru for Remote Cross Cultural Communication
Randy Gomez, Luis Merino
Robotic Intelligence
🎯 What it does: Proposed and implemented a prototype of a robot-mediated framework for cross-cultural communication, completed participatory design with interdisciplinary teams, identified robot roles and technical requirements, constructed a system prototype based on these requirements, and conducted pilot experiments with high school students in Japan and Australia, demonstrating that the system can stimulate students' interest in communicating, sharing, and discussing cultural topics with remote peers.
Design of Highly Repeatable and Multi-Functional Grippers for Precision Handling with Articulated Robots
Philip Gümbel, Klaus Dröder
Robotic Intelligence
🎯 What it does: Designed a low-cost, high-repeatability, multifunctional gripper optimized for joint robots to precisely handle components such as chip-level silicon wafers, achieving maximum repeatability during the grasping and releasing stages.
Design of Morphable StateNet Based on Pseudo-Generalization of Standing Up Motions for Humanoid with Variable Body Structure
Tasuku Makabe, Masayuki Inaba
Robotic Intelligence
🎯 What it does: Proposed a pseudo-generalized Morphable StateNet for standing-up motion of humanoid robots with variable body structures
Design of Two Morphing Robot Surfaces and Results from a User Study On What People Want and Expect of Them, Towards a "Robot-Room"
Nithesh Kumar, K. E. Green
Robotic Intelligence
🎯 What it does: Design and evaluate a deformable robot surface prototype, collecting user needs and expectations for the 'robot room'.
Design, Modeling and Analysis of a Spherical Parallel Continuum Manipulator for Nursing Robots
Zhenhua Gong, Ting Zhang
Robotic Intelligence
🎯 What it does: Designed, modeled, and analyzed a spherical parallel cable-driven continuous manipulator for nursing robots.
DESTINE: Dynamic Goal Queries with Temporal Transductive Alignment for Trajectory Prediction
Rezaul Karim, Amir Rasouli
Autonomous DrivingTransformerTime SeriesSequentialBenchmark
🎯 What it does: Proposed a trajectory prediction method based on dynamic target querying and temporal propagation alignment
Detecting and Mitigating System-Level Anomalies of Vision-Based Controllers
Aryaman Gupta, Somil Bansal
Anomaly DetectionAutonomous Driving
🎯 What it does: Proposes a runtime anomaly monitor that uses a reachability framework to perform offline stress testing on a visual controller, identifies system-level failures, trains a classifier online to label inputs that may lead to system failure, and designs a robust fallback controller to maintain system safety; validated on an autonomous aircraft landing system.
Development of a 3-RRS Micromanipulator Based on Origami-Inspired Spherical Joint
Haoqi Han, T. Arai
Robotic Intelligence
🎯 What it does: Developed a 3-RRS micro-actuator using origami-inspired spherical joints, and achieved miniaturization and low-cost design through PC-MEMS manufacturing process.
Development of an Automatic Sweet Pepper Harvesting Robot and Experimental Evaluation
Qinghui Pan, Chaochao Qiu
Pose EstimationRobotic IntelligenceImageAgriculture Related
🎯 What it does: Developed an automatic sweet pepper harvesting robot and conducted experimental validation in a plant factory; the main tasks include end-effector design, visual perception, and grasping posture control;
Development of the Assembling System for Structure Transformable Humanoid with Attach-Lock-Detachable Magnetic Coupling
Tasuku Makabe, Masayuki Inaba
Robotic Intelligence
🎯 What it does: Proposed a system that utilizes Attach‑Lock‑Detachable Magnetic Couplings (ALDMag) to achieve modular humanoid robot body structures, and validated its effectiveness through assembly experiments with small robots and full-scale arms.
Development of Variable Transmission Series Elastic Actuator for Hip Exoskeletons
Tianci Wang, Chunhua Liu
Robotic Intelligence
🎯 What it does: Developed a variable transmission series elastic actuator (VTSEA) for hip exoskeletons to meet torque-speed requirements under different assistance modes and achieve transparent human-robot interaction;
DexDLO: Learning Goal-Conditioned Dexterous Policy for Dynamic Manipulation of Deformable Linear Objects
Zhaole Sun, Robert B. Fisher
Robotic IntelligenceReinforcement Learning
🎯 What it does: Utilize a model-free reinforcement learning framework to learn dynamic dexterous manipulation of deformable linear objects (DLOs) using a fixed-base humanoid hand, abstracting various common DLO manipulation tasks as goal-conditioned tasks;
Dexterous In-hand Manipulation by Guiding Exploration with Simple Sub-skill Controllers
Gagan Khandate, M. Ciocarlie
Robotic IntelligenceReinforcement Learning
🎯 What it does: Utilizing simple sub-skill controllers from domain knowledge to guide exploration, thereby improving sample efficiency in learning palm manipulation within simulated environments
Dexterous Legged Locomotion in Confined 3D Spaces with Reinforcement Learning
Zifan Xu, Peter Stone
Robotic IntelligenceReinforcement Learning
🎯 What it does: Propose a hierarchical locomotion controller that integrates classical path planning with a reinforcement learning-based low-level policy for goal-oriented quadruped robot navigation in narrow three-dimensional spaces.
Differentiable Boustrophedon Paths That Enable Optimization Via Gradient Descent
Thomas Manzini, Robin R. Murphy
OptimizationComputational EfficiencyRepresentation Learning
🎯 What it does: Proposed a differentiable B-spline path representation, explored additional optimizable parameters, and experimentally validated that this representation can high-fidelity reconstruct the scores of traditional discrete representations; attempted to optimize using gradient descent but failed due to the highly non-convex search space.
Differentiable Compliant Contact Primitives for Estimation and Model Predictive Control
Kevin Haninger, L. Roveda
OptimizationRobotic Intelligence
🎯 What it does: Proposed a framework based on differentiable compliant contact primitives for online estimation of contact models and applying them to model predictive control (MPC);
Differentially Encoded Observation Spaces for Perceptive Reinforcement Learning
Lev Grossman, B. Plancher
CompressionReinforcement LearningImageVideoBenchmark
🎯 What it does: Using differential encoding compressed sensing to compress the experience replay buffer of deep reinforcement learning, and fully load it into RAM without affecting training performance
Diffusion-Based Point Cloud Super-Resolution for mmWave Radar Data
Kai-Rui Luan, Xieyuanli Chen
Super ResolutionDiffusion modelPoint CloudStochastic Differential Equation
🎯 What it does: Proposed a 3D millimeter-wave radar point cloud super-resolution method called Radar-diffusion.
DINOBot: Robot Manipulation via Retrieval and Alignment with Vision Foundation Models
Norman Di Palo, Edward Johns
RetrievalRobotic IntelligenceTransformerContrastive LearningImage
🎯 What it does: Proposes DINOBot, a robotic imitation learning framework for robotic manipulation that utilizes DINO-trained Vision Transformer features for image-level and pixel-level retrieval and alignment.
DiPPeR: Diffusion-based 2D Path Planner applied on Legged Robots
Jianwei Liu, Dimitrios Kanoulas
Robotic IntelligenceConvolutional Neural NetworkDiffusion modelImage
🎯 What it does: Proposed DiPPeR, a diffusion-based fast 2D path planning framework, validated on mazes and real robots (Spot, Go1).
Direct 3D model-based object tracking with event camera by motion interpolation
Yufan Kang, Takeshi Oishi
Object TrackingOptical Flow
🎯 What it does: Proposes a method for object tracking based on 3D models using events captured directly by an event camera.
Direct learning of home vector direction for insect-inspired robot navigation
Michiel Firlefyn (Delft University of Technology), Guido C. H. E. de Croon (Delft University of Technology)
Robotic IntelligenceConvolutional Neural NetworkSimultaneous Localization and MappingImage
🎯 What it does: Studied a vision-based homing vector learning and navigation method, which utilizes a convolutional neural network (CNN) to directly learn the home direction from visual perception during the learning flight phase, and subsequently eliminates drift during the return phase through ranging and visual inference.
Direct Self-Identification of Inverse Jacobians for Dexterous Manipulation Through Particle Filtering
Joshua T. Grace, A. Dollar
Robotic Intelligence
🎯 What it does: Propose a method that utilizes a self-identified inverse Jacobian matrix to control robotic in-hand manipulation.
Directly 3D Printed, Pneumatically Actuated Multi-Material Robotic Hand
Hanna Matusik, Daniela Rus
Robotic Intelligence
🎯 What it does: Designed and manufactured a soft robotic hand that can be directly 3D printed, composed of multiple materials, featuring 15 degrees of freedom, and includes five fingers with a thumb.
Discovering Biological Hotspots with a Passively Listening AUV
Seth McCammon, Y. Girdhar
MultimodalityAudio
🎯 What it does: Developed a passive listening AUV system that integrates multiple sensing modalities for audio-visual surveys in coral reef areas, aiding marine biologists in collecting data to understand the ecological relationships between fish and other organisms and their habitats.
Discuss Before Moving: Visual Language Navigation via Multi-expert Discussions
Yuxing Long, Hao Dong
Robotic IntelligenceLarge Language ModelMixture of ExpertsVision Language ModelImageText
🎯 What it does: Designed a zero-shot vision-language navigation framework called DiscussNav, allowing the navigation agent to actively discuss with multi-domain experts before each movement step to acquire critical information;
DISO: Direct Imaging Sonar Odometry
S. Xu, Sen Wang
Simultaneous Localization and MappingImage
🎯 What it does: Developed a direct imaging sonar odometry system called DISO for estimating the relative spatial transformation between two sonar image frames.
Distill-then-prune: An Efficient Compression Framework for Real-time Stereo Matching Network on Edge Devices
Baiyu Pan, Jun Cheng
Depth EstimationKnowledge DistillationImage
🎯 What it does: Propose a lightweight stereo matching network compression framework that combines knowledge distillation and model pruning, achieving both real-time performance and high accuracy.
Distilling and Retrieving Generalizable Knowledge for Robot Manipulation via Language Corrections
Lihan Zha (Stanford University), Dorsa Sadigh (Stanford University)
Robotic IntelligenceLarge Language ModelImageText
🎯 What it does: Proposes a system called DROC based on large language models (LLMs) for online processing of any form of language feedback, extracting general knowledge and retrieving relevant experiences to enhance robot performance in new environments.
Distributed Control Barrier Functions for Global Connectivity Maintenance
N. D. Carli, P. Giordano
Optimization
🎯 What it does: Proposes a distributed implementation of a quadratic programming (QP) controller framework that utilizes control barrier functions (CBF) to maintain global connectivity among multiple quadrotors under communication and perception constraints;
Distributed Multi-robot Online Sampling with Budget Constraints
Azin Shamshirgaran, Stefano Carpin
OptimizationRobotic Intelligence
🎯 What it does: Proposes an online distributed multi-robot sampling algorithm based on Monte Carlo Tree Search (MCTS), planning sampling paths for each robot under a budget constraint
Distribution-Aware Continual Test-Time Adaptation for Semantic Segmentation
Jiayin Ni, Shanghang Zhang
SegmentationDomain AdaptationImage
🎯 What it does: Propose a distribution-aware tuning (DAT) method to achieve efficient and feasible continual test-time adaptation (CTTA) for semantic segmentation tasks
Distributional Reinforcement Learning with Sample-set Bellman Update
Weijia Zhang, Yang Yu
Reinforcement LearningVideo
🎯 What it does: Propose a distributed reinforcement learning framework that accurately models the reward distribution using a Gaussian Mixture Model.
Distributionally Robust Chance Constrained Trajectory Optimization for Mobile Robots within Uncertain Safe Corridor
Shaohang Xu, Chin Pang Ho
OptimizationRobotic Intelligence
🎯 What it does: This paper proposes distributionally robust safe corridor constraints (DRSCCs) and integrates them into a trajectory optimization framework based on Bernstein basis functions, proving that they can be transformed into a convex quadratic program.
Distributionally Robust CVaR-Based Safety Filtering for Motion Planning in Uncertain Environments
Sleiman Safaoui, Tyler H. Summers
Autonomous DrivingOptimizationSafty and PrivacyTime Series
🎯 What it does: Proposed a computationally efficient safety filtering scheme that reduces the collision risk of the ego vehicle's motion planning by utilizing multiple obstacle trajectory prediction samples.
Dive Deeper into Rectifying Homography for Stereo Camera Online Self-Calibration
Hongbo Zhao, Rui Fan
Pose EstimationOptimizationRectified FlowVideo
🎯 What it does: Propose an online self-calibration algorithm for stereo cameras based on corrected single-lens single-image pairs, present a globally optimal external parameter estimation method in stereo video sequences, and introduce four new evaluation metrics to measure the robustness and accuracy of external parameter estimation.
Diving into the Depths of Spotting Text in Multi-Domain Noisy Scenes
Alloy Das, Josep Llad'os
Object DetectionSuper ResolutionDomain AdaptationTransformerImageBenchmark
🎯 What it does: Studied and proposed a domain-agnostic scene text detection system called DA-TextSpotter, and created a benchmark for underwater noisy scenarios named UWT.
DL-PoseNet: A Differential Lightweight Network for Pose Regression over SE(3)
Wenjie Li, Lijun Chen
Pose EstimationImage
🎯 What it does: Propose DL-PoseNet, a lightweight deep network for SE(3) pose regression;
DMSA - Dense Multi Scan Adjustment for LiDAR Inertial Odometry and Global Optimization
D. Skuddis, Norbert Haala
Autonomous DrivingOptimizationSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Proposed the DMSA method to achieve dense joint registration of multi-frame point clouds, using full point cloud merging and gradually reducing scattering through a Gaussian distribution model based on uniform grid cells; combined with IMU measurements for sliding window continuous trajectory optimization and large-scale keyframe optimization.
Do we need scan-matching in radar odometry?
V. Kubelka, Martin Magnusson
Autonomous DrivingSimultaneous Localization and MappingPoint Cloud
🎯 What it does: Compared various odometry estimation methods based on 4D radar, including direct integration of Doppler/IMU data, Kalman filter fusion, and 3D point cloud scan alignment, and evaluated their performance.
Do We Run Large-scale Multi-Robot Systems on the Edge? More Evidence for Two-Phase Performance in System Size Scaling
Jonas Kuckling, H. Hamann
OptimizationComputational EfficiencyRobotic Intelligence
🎯 What it does: Explored the performance of multi-robot systems when scaling up, finding that at the critical scale, system performance splits into two stages: near-optimal and worst-case.
Doduo: Learning Dense Visual Correspondence from Unsupervised Semantic-Aware Flow
Zhenyu Jiang (University of Texas at Austin), Yuke Zhu (University of Texas at Austin)
Flow-based ModelOptical FlowImage
🎯 What it does: Proposes the Doduo method for learning dense visual correspondence from unsupervised flow with semantic priors.
Domain Adaptation of Visual Policies with a Single Demonstration
Weiyao Wang, Gregory D. Hager
Domain AdaptationTransformerPrompt Engineering
🎯 What it does: Propose the PromptAdapt framework, which utilizes a single demonstration (prompt) to train a visual strategy, achieving adaptation in the target domain
Domain Randomization for Sim2real Transfer of Automatically Generated Grasping Datasets
J. Huber, Stéphane Doncieux
Domain AdaptationRobotic Intelligence
🎯 What it does: This paper generated and tested more than 7,000 grasping trajectories using quality diversity (QD) methods on three robotic arms and grippers, exploring the feasibility of automatically generated grasping datasets on real robots.
Doppler-only Single-scan 3D Vehicle Odometry
Andres Galeote-Luque, Javier González Jiménez
Autonomous Driving
🎯 What it does: Estimate the complete three-dimensional motion of a vehicle using only a single scan from a Doppler sensor.
DOS®: A Deployment Operating System for Robots
Guo Ye, Han Liu
Robotic Intelligence
🎯 What it does: Proposed a robot deployment operating system named DOS and verified its reliability in both production and simulation environments.
Dream2Real: Zero-Shot 3D Object Rearrangement with Vision-Language Models
Ivan Kapelyukh, Edward Johns
Robotic IntelligenceImageText
🎯 What it does: Propose the Dream2Real framework, integrating vision-language models (VLMs) trained on 2D data into the 3D object rearrangement process. The robot autonomously constructs a 3D scene, virtually rearranges objects, renders images, and then selects the arrangement most consistent with user instructions through VLM evaluation. The selected arrangement is subsequently implemented in the real world using grasp-and-place operations.
Drive Anywhere: Generalizable End-to-end Autonomous Driving with Multi-modal Foundation Models
Tsun-Hsuan Wang, Daniela Rus
Autonomous DrivingTransformerVision Language ModelWorld ModelMultimodality
🎯 What it does: Built a generalizable end-to-end autonomous driving system based on a multimodal foundation model, leveraging Transformers to extract fine-grained spatial features, employing latent space simulation for training and strategy debugging, and extending the model's capabilities through pixel/patch-aligned feature descriptors. The system enhances robustness to out-of-distribution scenarios by fusing language and visual perception.
DRIVE: Data-driven Robot Input Vector Exploration
D. Baril, François Pomerleau
Autonomous DrivingComputational EfficiencyData-Centric LearningRobotic Intelligence
🎯 What it does: Proposes a data-driven robot input vector exploration (DRIVE) protocol to collect input limits for unmanned ground vehicles (UGV) and gather training data for motion models, while introducing a slippage learning model that outperforms traditional acceleration-based learning methods.
Driving with LLMs: Fusing Object-Level Vector Modality for Explainable Autonomous Driving
Long Chen, Jamie Shotton
Autonomous DrivingExplainability and InterpretabilityTransformerLarge Language ModelReinforcement LearningMultimodality
🎯 What it does: Propose an object-level multimodal LLM architecture that integrates vectorized numerical modalities with pre-trained LLMs to enhance driving scene understanding and constructs a 160k QA pair dataset.
DroneMOT: Drone-based Multi-Object Tracking Considering Detection Difficulties and Simultaneous Moving of Drones and Objects
Peng Wang, De-qin Li
Object TrackingVideo
🎯 What it does: Proposed a multi-object tracking method for UAV scenarios called DroneMOT, including dual-domain integrated attention module, motion-driven association scheme, adaptive feature synchronization technology, and dual motion prediction methods, aiming to enhance detection and association performance in environments with small targets, blurriness, and occlusions.
DTPP: Differentiable Joint Conditional Prediction and Cost Evaluation for Tree Policy Planning in Autonomous Driving
Zhiyu Huang, Chen Lv
Autonomous DrivingTransformer
🎯 What it does: This paper proposes a tree-structured policy planning approach and designs a differentiable joint training framework that integrates self-vehicle conditional motion prediction with cost assessment, directly enhancing the final planning performance.
Dual Quaternion Control of UAVs with Cable-suspended Load
Yuxia Yuan, Markus Ryll
OptimizationRobotic Intelligence
🎯 What it does: Unify the dynamics and kinematics modeling and control of UAV suspended loads using dual quaternions to solve the lifting and tracking problems of hanging loads
Dual-Critic Deep Reinforcement Learning for Push-Grasping Synergy in Cluttered Environment
Jia-Xing Zhong, Xiaoqi Chen
OptimizationRobotic IntelligenceReinforcement LearningImage
🎯 What it does: Proposed a dual evaluator deep reinforcement learning framework to optimize push-pull collaborative grasping in dense cluttered environments, reducing pre-grasping redundancy.
Dual-IMU State Estimation for Relative Localization of Two Mobile Agents
Wenqian Lai, Kejian J. Wu
Autonomous DrivingSimultaneous Localization and Mapping
🎯 What it does: Studied the relative positioning problem of two moving bodies under a dual IMU system, and analyzed the system's dynamic model and observability
Dual-modal Tactile E-skin: Enabling Bidirectional Human-Robot Interaction via Integrated Tactile Perception and Feedback
Shilong Mu, Wenbo Ding
Robotic IntelligenceMultimodality
🎯 What it does: Propose a dual-mode electronic skin that integrates magnetic tactile sensing and vibration feedback to enhance human-computer interaction experience
DualAT: Dual Attention Transformer for End-to-End Autonomous Driving
Zesong Chen, Xiaojun Tan
Autonomous DrivingTransformerMultimodalityBenchmark
🎯 What it does: Proposes a multi-task imitation learning framework that utilizes a dual-attention Transformer to enhance multi-modal fusion and waypoint prediction;
Dusk Till Dawn: Self-supervised Nighttime Stereo Depth Estimation using Visual Foundation Models
M. Vankadari, Niki Trigoni
Depth EstimationTransformerImage
🎯 What it does: Developed a self-supervised stereo depth estimation algorithm for nighttime scenes, leveraging a pre-trained visual foundation model to extract general features and combining a novel masking method to filter pixels violating photometric consistency.