arXivSub Start free trial

IROS 2025 Papers — Page 6

IEEE/RSJ International Conference on Intelligent Robots and Systems · 1984 papers

Design of an Affordable, Fully-Actuated Biomimetic Hand for Dexterous Teleoperation Systems

Zhaoliang Wan, Hui Cheng

Robotic Intelligence

🎯 What it does: Developed a low-cost, fully actuated 5-fingered biomimetic hand (RAPID Hand) for an enhanced teleoperation system.

Design of An Electromagnetically Modulated Resistance Mechanism to Realize Compact Passive Force-Feedback Wearable Devices

Rene M. Suarez Flores, S. Nisar

Physics Related

🎯 What it does: Designed and verified a passive force feedback wearable device based on dual electromagnets for dynamic adjustment of tendon tension.

Design of DNA Origami-Engineered Tetrahedral Nanorobots

Haowen Chen, Xiaoming Liu

Robotic IntelligenceDrug DiscoveryBiomedical Data

🎯 What it does: Studied reconfigurable tetrahedral DNA nanobots and verified their conformational transition paths using multi-resolution molecular dynamics; systematically analyzed their structural stability, fabricated molecule-triggered high-stability, high-drug-delivery-capability tetrahedral DNA nanobots, and applied them for identifying and inhibiting circulating tumor cells.

Design of Q8bot: A Miniature, Low-Cost, Dynamic Quadruped Built with Zero Wires

Yufeng Wu, Dennis Hong

Robotic Intelligence

🎯 What it does: Introduces Q8bot 2, an open-source, miniature quadruped robot designed to provide a low-cost, easily reproducible platform for robotics research and education;

Design of scalable orthogonal digital encoding architecture for large-area flexible tactile sensing in robotics

Weijie Liu, Yancheng Wang

Robotic Intelligence

🎯 What it does: Designed a scalable orthogonal digital encoding architecture for large-area flexible tactile sensing, utilizing distributed nodes to achieve parallel superposition of energy-orthogonal basis codes, significantly reducing wiring requirements and enhancing data throughput.

Design Optimization of a Single-DoF Gait Rehabilitation Robot for a Domestic Environment

Julius Ambros, Sami Haddadin

OptimizationRobotic Intelligence

🎯 What it does: Proposed a single-degree-of-freedom chain-driven gait rehabilitation robot suitable for home environments, and optimized its design parameters.

Design Optimization of Three-Dimensional Wire Arrangement Considering Wire Crossings for Tendon-driven Robots

Kento Kawaharazuka, Kei Okada

OptimizationRobotic Intelligence

🎯 What it does: Studied the wire layout optimization problem in three-dimensional joint-driven robots considering wire crossing;

Design, Manufacturing, and Experiments of an Origami-based Parallel-Legged Structure for Insect-scale Robots

Qunwei Zhu, Zirong Luo

Robotic Intelligence

🎯 What it does: Designed, manufactured, and tested a foldable parallel leg structure based on origami, integrating robot hinges, links, and actuators using multi-layer composite laminates to achieve rapid assembly of insect-scale robots.

Designing a Magnetic Endoscope for In Vivo Contact-Based Tissue Scanning Using Developable Roller

N. Greenidge, Pietro Valdastri

Robotic IntelligenceBiomedical Data

🎯 What it does: This paper designs a magnetic endoscope utilizing an expandable roller (oloid shape), achieving rolling motion under magnetic manipulation, equipped with a contact sensor, enabling scanning of the colon wall in in vivo experiments.

Detecting Obstacles on Railroads Using Computer Vision on UAVs

Aryan Anand, Nikolaos Vitzilaios

Object DetectionImage

🎯 What it does: Proposed an object detection system that can be implemented on drones for detecting obstacles on railways

Detection for Harvesting with an Active Illumination Camera System and DUTU2-Net+

Qinghui Pan, Dong Wang

Object DetectionConvolutional Neural NetworkImageAgriculture Related

🎯 What it does: A lightweight sweet pepper handle detection method is proposed, combining an active illumination camera system with DUTU2-Net+ to achieve fast and accurate handle localization.

Development and Characterization of an Adaptive Baromorphic End-Effector for Precision Agricultural Handling

D. Cafolla, M. I. N. Al Khatib

Robotic IntelligenceAgriculture Related

🎯 What it does: Developed an adjustable gripper end-effector equipped with sensors for precision agricultural handling

Development of a 4-DOF Mobile Manipulator for Repetitive Gait Training on the Track for Stroke Patients

Junyeong Lee, Jungwon Yoon

Robotic Intelligence

🎯 What it does: Developed a 4-degree-of-freedom mobile manipulator for repetitive gait training of stroke patients on a track.

Development of a Cleaning Robot Capable of Self-Propelled Cleaning for Ducts in Real-World Environments Employing a Planetary Gear Mechanism

Y. Ono, Taro Nakamura

Robotic Intelligence

🎯 What it does: Developed a self-propelled cleaning robot capable of removing accumulated grease from restaurant kitchen pipelines and operating autonomously in environments inaccessible to humans.

Development of a hard matter crushing peristaltic bioreactor inspired by an avian gizzard structure for Fermentation Acceleration

K. Kikyodani, Taro Nakamura

🎯 What it does: Developed a rhythmic bioreactor inspired by the structure of a bird's stomach, featuring enhanced fragmentation capability with fixed and movable spherical solids;

Development of a Novel Miniaturized Dexterous Manipulator with Variable Stiffness for NOTES

Rong Cong, Changsheng Li

Robotic IntelligenceBiomedical Data

🎯 What it does: Designed and verified a 5 mm diameter, 7 degrees of freedom variable stiffness small flexible manipulator for NOTES

Development of a Soft Robotic Fish with Stiffness Modulation and Wriggling Locomotion

Hua Shao, Fengran Xie

Robotic Intelligence

🎯 What it does: This paper designs and implements a soft robotic fish with adjustable stiffness and peristaltic movement;

Development of an Efficient Stiffness Modulation Mechanism in Fish-like Robots for Enhanced Swimming Performance

Xu Chao, Xingjian Jing

Robotic Intelligence

🎯 What it does: Studied an online stiffness regulation mechanism for the tail of a fish-shaped robot, which can achieve stiffness regulation without the need for additional actuators or power sources to enhance swimming performance.

Development of an under-actuated tendon-driven planar elephant robot based on synergistic motion analysis

Koki Kitabayashi, Ryuta Ozawa

Robotic IntelligenceVideo

🎯 What it does: This paper develops an underactuated tendon-driven planar elephant trunk robot based on collaborative motion analysis. First, principal components of the trunk's motion are extracted through video data analysis, then a tendon-driven transmission system and control system are designed to achieve collaborative motion with limited actuators, and the motion accuracy is evaluated.

Development of Variable Chain Motor with Shape and Speed-Torque Characteristics Variability and Its Application to a Humanoid

Hiromi Tada, Kei Okada

Robotic Intelligence

🎯 What it does: Developed and evaluated the Variable Chain Motor (VC Motor), applying it to the elbow joint of a humanoid robot arm to achieve high-speed, high-load operations.

Development of Wearable Assistive Robots Using Artificial Muscle for Older Adults

Yafei Zhao, Ning Xi

Robotic IntelligenceBiomedical Data

🎯 What it does: Developed a wearable robotic assistant system that uses artificial muscle actuators to directly compensate for muscle weakness in the elderly, improving their balance and mobility functions.

DexPour: Effective and Efficient High-DoF Robotic Hand Liquid Pouring via Hierarchical Reward with Approximated Proxy Abstraction

Xinmin Fang, Zhengxiong Li

Computational EfficiencyRobotic IntelligenceReinforcement Learning

🎯 What it does: Propose a reinforcement learning method called DexPour, combining hierarchical rewards and approximate proxy abstraction (APA), to achieve liquid pouring tasks with high-degree-of-freedom robotic hands;

DEXTER-LLM: Dynamic and Explainable Coordination of Multi-Robot Systems in Unknown Environments via Large Language Models

Yuxiao Zhu, Zhongkui Li

Explainability and InterpretabilityRobotic IntelligenceLarge Language ModelTextChain-of-Thought

🎯 What it does: Proposed the DEXTER-LLM framework to achieve dynamic task planning for multi-robot systems in unknown environments, leveraging LLM for task decomposition and explanation;

Dexterous Manipulation Based on Prior Dexterous Grasp Pose Knowledge

Hengxu Yan, Cewu Lu

Robotic IntelligenceReinforcement Learning

🎯 What it does: Proposes a reinforcement learning method that leverages prior grasp pose knowledge, dividing the operation process into two stages: grasp pose generation and environment exploration;

DG16M: A Large-Scale Dataset for Dual-Arm Grasping with Force-Optimized Grasps

Md Faizal Karim, Equal Contribution

Robotic IntelligenceBenchmark

🎯 What it does: Proposed a dataset containing 16 million dual-arm grasps and created a benchmark dataset for evaluating grasp quality;

DGETP: Dynamic Graph Attention Network for Embodied Task Planning

Pengfei Sun, Jian Yu

Robotic IntelligenceGraph Neural NetworkGraphSequential

🎯 What it does: Propose a Dynamic Graph Attention Network (DGETP) for embedded task planning, capable of processing scene graph sequences and robot graphs to capture environmental dynamics;

DGVO: A Dynamically Constrained Gradient Velocity Obstacle Approach for Mobile Robots in Dynamic Environments

Bowen Xiao, Ruocheng Li

Autonomous DrivingOptimization

🎯 What it does: A framework based on velocity obstacles is proposed to address obstacle avoidance for constrained mobile robots in dynamic environments. By defining velocity feasible regions (VFR) and dynamic constraint gradient velocity obstacles (DGVO) through nonlinear mapping, collision-safe velocities are achieved in real-time via unconstrained gradient descent optimization.

DHC-ME: A Decentralized Hybrid Cooperative Approach for Multi-Robot Autonomous Exploration

Wenhao Jia, Liang Li

Robotic IntelligenceAgentic AISimultaneous Localization and Mapping

🎯 What it does: Proposed and implemented DHC-ME: a decentralized hybrid collaboration strategy for multi-robot range-aware exploration, enhancing team coordination and exploration efficiency.

Diegetic Graphical User Interfaces for Robot Control via Eye-gaze

E. N. Sardinha, V. Garate

Robotic IntelligenceBenchmark

🎯 What it does: Designed and implemented an eye-tracking-based visual interactive interface for precision pick-and-place tasks on robotic arms, evaluated with 21 participants

Diff-IP2D: Diffusion-Based Hand-Object Interaction Prediction on Egocentric Videos

Junyi Ma, Hesheng Wang

Diffusion modelVideo

🎯 What it does: Propose Diff-IP2D, a non-autoregressive bidirectional constrained hand-object interaction prediction method based on diffusion.

Diff-MSM: Differentiable MusculoSkeletal Model for Simultaneous Identification of Human Muscle and Bone Parameters

Yingfan Zhou, Cheng Fang

Biomedical Data

🎯 What it does: Proposed and implemented a differentiable musculoskeletal model (Diff-MSM), which uses end-to-end automatic differentiation to jointly identify muscle and skeletal parameters from measurable muscle activation signals to observable motion trajectories, without requiring measurement of internal joint torques.

Differentiable Rendering-based Pose Estimation for Surgical Robotic Instruments

Zekai Liang, Michael C. Yip

Pose EstimationBiomedical Data

🎯 What it does: This paper proposes a surgical robot tool pose estimation method based on differentiable rendering, which constructs a pose matching pipeline using geometric primitives (e.g., cylinders) to achieve markerless initialization with single calibration;

Differential Six-Axis Force and Torque Measurement in a Prototype Robotic Surgical Instrument

Daniel Neykov, Sebastian Matich

Robotic IntelligenceBiomedical Data

🎯 What it does: Propose using two small six-axis F/T sensors (proximal and distal) for differential measurement to cancel out interference caused by cable-driven systems in robotic surgical instruments, and validate the effectiveness on an experimental cable-driven forceps.

Differential-Flatness-Based Tracking Control for Tractor-Trailers in Reversing Maneuvers

Bo Yang, Wen Yang

Autonomous Driving

🎯 What it does: This paper proposes a controller based on differential flatness (DFBC) to achieve precise trajectory tracking for a tractor-trailer combination during reversing.

DiffGen: Robot Demonstration Generation via Differentiable Physics Simulation, Differentiable Rendering, and Vision-Language Model

Yang Jin, Cewu Lu

GenerationRobotic IntelligenceVision Language ModelMultimodalityPhysics Related

🎯 What it does: Automatically and efficiently generate robot demonstrations through differentiable physics simulation, differentiable rendering, and vision-language models

DiffSSC: Semantic LiDAR Scan Completion using Denoising Diffusion Probabilistic Models

Helin Cao, Sven Behnke

Autonomous DrivingDiffusion modelPoint Cloud

🎯 What it does: Complete semantic LiDAR point cloud scene using a denoising diffusion probabilistic model.

Diffusion Policies for Risk-Averse Behavior Modeling in Offline Reinforcement Learning

Xiaocong Chen, Lina Yao

Reinforcement LearningDiffusion model

🎯 What it does: Propose a risk-averse distributed offline reinforcement learning method that can simultaneously consider epistemic uncertainty and environmental stochasticity.

Diffusion Policies with Value-Conditional Optimization for Offline Reinforcement Learning

Yunchang Ma, Xin Xu

OptimizationReinforcement LearningDiffusion modelBenchmark

🎯 What it does: Propose a novel offline reinforcement learning method called DIVO, which generates high-quality in-distribution state-action samples using diffusion models and improves the policy through value-conditioned optimization.

Diffusion Suction Grasping with Large-Scale Parcel Dataset

Ding-Tao Huang, Long Zeng

GenerationData SynthesisRobotic IntelligenceDiffusion modelPoint CloudBenchmark

🎯 What it does: Created the Parcel-Suction-Dataset and proposed the Diffusion-Suction framework

Diffusion-Based Approximate MPC: Fast and Consistent Imitation of Multi-Modal Action Distributions

Pau Marquez Julbe, K. Kuchenbecker

OptimizationRobotic IntelligenceDiffusion model

🎯 What it does: Approximating model predictive control (MPC) using diffusion models, achieving a fast and consistent imitation learning controller that operates in joint space at high sampling rates (up to kilohertz).

Diffusion-FS: Multimodal Free-Space Prediction via Diffusion for Autonomous Driving

Keshav Gupta, Madhava Krishna

Autonomous DrivingDiffusion modelImage

🎯 What it does: Propose to treat predictable drivable free-space corridor prediction as a pure image perception task using monocular camera inputs; develop a self-supervised method to generate free-space samples through future vehicle trajectories and front-facing images; model the corridor distribution in images using diffusion processes; introduce the ContourDiff architecture, achieving structured and interpretable corridor prediction through denoising of contour points.

DiFuse-Net: RGB and Dual-Pixel Depth Estimation using Window Bi-directional Parallax Attention and Cross-modal Transfer Learning

Kunal Swami, Pankaj Bajpai

Depth EstimationImageMultimodality

🎯 What it does: Designed and implemented DiFuse-Net for RGB and dual-pixel (DP) depth estimation, and proposed a cross-modal transfer learning (CmTL) mechanism as well as a new Dual-Camera Dual-Pixel (DCDP) dataset.

Direct, Targetless and Automatic Joint Calibration of LiDAR-Camera Intrinsic and Extrinsic

Yishu Shen, Tong Qin

Pose EstimationAutonomous DrivingOptimizationImagePoint Cloud

🎯 What it does: A direct, target-free, and automatic LiDAR-camera joint calibration method is proposed, which employs a two-stage alternating optimization to simultaneously improve intrinsic and extrinsic parameters.

Directed Spatial Consistency-Based Partial-to-Partial Point Cloud Registration with Deep Graph Matching

Jingwen Zhou, Zhe Min

Pose EstimationGraph Neural NetworkPoint CloudBiomedical Data

🎯 What it does: Propose a partial-to-partial point cloud registration framework based on directional space consistency, which first extracts overlapping regions and obtains a hard matching matrix via graph matching, then generates translation-invariant edge vectors through sampling nodes, combines point-level and edge-level geometric constraints for dual optimization, and introduces a bidirectional registration mechanism to enhance registration stability.

Disambiguate Gripper State in Grasp-Based Tasks: Pseudo-Tactile as Feedback Enables Pure Simulation Learning

Yifei Yang, Yue Wang

Robotic Intelligence

🎯 What it does: Propose utilizing pseudo-haptic feedback (tactile sensing realized through force-controlled grippers) to eliminate ambiguity in gripper states, thereby enhancing the robustness of imitation learning strategies for grasping tasks and achieving pure simulation learning;

DISCOVERSE: Efficient Robot Simulation in Complex High-Fidelity Environments

Yufei Jia, Guyue Zhou

Domain AdaptationRobotic IntelligenceGaussian Splatting

🎯 What it does: Proposed Discoverse, a unified, modular, open-source 3D simulation framework based on 3DGS for Real2Sim2Real robot learning;

Disentangled Object-Centric Image Representation for Robotic Manipulation

David Emukpere, Seungsu Kim

Representation LearningRobotic IntelligenceImage

🎯 What it does: Proposed a object-centric framework named DOCIR, which introduces decoupled representations for objects of interest, obstacles, and the robot's body to learn robotic manipulation skills from visual inputs;

Disentangling Uncertainty for Safe Social Navigation using Deep Reinforcement Learning

Daniel Flögel, Sören Hohmann

Autonomous DrivingReinforcement Learning

🎯 What it does: Integrate model uncertainty (including aleatoric, epistemic, and predictive) into a deep reinforcement learning framework to achieve safe social navigation.

Distance and Collision Probability Estimation from Gaussian Surface Models

K. Goel, Wennie Tabib

Computational EfficiencyRobotic IntelligenceMesh

🎯 What it does: Propose a method that uses a Gaussian Mixture Model to represent the environment and employs an elliptical robot with an elliptical approximation of the surface model to compute the collision probability, Euclidean distance, and gradient between the robot and the surface without explicitly constructing a free space representation.

Distillation-PPO: A Novel Two-Stage Reinforcement Learning Framework for Humanoid Robot Perceptive Locomotion

Qiang Zhang, Renjing Xu

Knowledge DistillationRobotic IntelligenceReinforcement Learning

🎯 What it does: Proposed a novel two-stage reinforcement learning framework that combines a teacher policy trained in a fully observable MDP with a student policy trained in a POMDP, using the teacher policy to regularize and supervise the student policy for achieving human-like robotic perception-based walking control.

Distilling 3D distinctive local descriptors for 6D pose estimation

Amir Hamza, Fabio Poiesi

Pose EstimationKnowledge DistillationImage

🎯 What it does: Train a student model using a knowledge distillation framework to regress 3D local descriptors from the GeDi teacher, thereby improving efficiency

Distilling Realizable Students from Unrealizable Teachers

Yujin Kim, Sanjiban Choudhury

Knowledge DistillationReinforcement Learning

🎯 What it does: Proposes a policy distillation method under privileged information, allowing students with partial observations to learn from teachers with full state access, and enhances learning effectiveness through active interaction with the teacher.

Distributed Autonomous Safe Flight Planning for Multiple UAVs in Unknown Environments

Fan Yang, Youngjin Choi

OptimizationSafty and Privacy

🎯 What it does: Proposes two techniques: optimizing the front-end path generated by traditional path planning to match UAV dynamics, resulting in the back-end motion trajectory; introducing a collision detection adjustment area and achieving dynamic trajectory re-planning for multiple UAVs through local neighborhood communication, thereby enabling collision avoidance.

Distributed Contact Sensing Enabled by Vibration Propagation on Robot End-Effector

W. Tan, Yitian Shao

Robotic IntelligenceTime Series

🎯 What it does: Designed a tactile perception system that uses a single accelerometer and transmits vibration signals through a string to detect remote contact position, contact force, and surface material of the object in contact on the robot end-effector.

Distributed Cooperative Target Tracking and Active Sensing of Dual-AUV Based on Flank Array Sonar Detection

Qi Qi, Yanjie Pan

Object TrackingAudio

🎯 What it does: Proposes a distributed collaborative tracking and active sensing scheme for a dual AUV system based on side-lobe array sonar

Distributed Fault-Tolerant Multi-Robot Cooperative Localization in Adversarial Environments

Tohid Kargar Tasooji, Ramviyas Parasuraman

Robotic IntelligenceSimultaneous Localization and Mapping

🎯 What it does: Propose a distributed fault-tolerant multi-robot collaborative localization framework combined with an adaptive event-triggered communication strategy to enhance robustness against sensor failures and communication disruptions in adversarial environments;

Distributed Oscillatory Guidance for Formation Flight of Fixed-Wing Drones

Yang Xu, Héctor García de Marina

Robotic Intelligence

🎯 What it does: Proposed a distributed oscillation guidance algorithm to enable fixed-wing UAVs to achieve formation flight without altering their speed, utilizing oscillatory behavior to regulate average path velocity and achieving adaptive oscillation amplitude through neighbor communication.

Distributional Decision Transformer: Risk-Sensitive Offline RL via Quantile-Based Critics and Stochastic Return

Changxu Wei, Wenbo Ding

TransformerReinforcement LearningSequential

🎯 What it does: Proposed the Distributional Decision Transformer (DDT), an offline reinforcement learning framework that integrates probabilistic return distribution modeling with autoregressive action generation.

Disturbance Observer-based Control Barrier Functions with Residual Model Learning for Safe Reinforcement Learning

Dvij Kalaria, John M. Dolan

Autonomous DrivingReinforcement LearningBenchmark

🎯 What it does: Proposes a safety reinforcement learning framework based on robust control barrier functions (CBF), which quantifies the difference between the nominal model and actual dynamics by combining disturbance observer and residual model learning, and validates it on Safety-gym benchmarks, car robots, and real-world F1/10 racing.

Diverse and Adaptive Behavior Curriculum for Autonomous Driving: A Student-Teacher Framework with Multi-Agent RL

A. Abouelazm, J. M. Zöllner

Autonomous DrivingGraph Neural NetworkReinforcement Learning

🎯 What it does: Propose a student-teacher framework where a graph-based multi-agent RL teacher automatically generates a multi-difficulty traffic behavior curriculum, training a deep RL student to complete driving tasks in partially observable environments.

DL-Clip: Online D-Learning with Clipping Operation for Fast Model-Free Stabilizing Control

Jingxuan Liu, Quan Quan

OptimizationReinforcement Learning

🎯 What it does: Propose DL-Clip, an online learning method for nonlinear stable control, capable of functioning without prior knowledge of system dynamics and reward signals, while significantly improving training efficiency.

DMPBot: A high-speed, high-precision, omnidirectional, insect-scale piezoelectric robot

Yan Chen, Xiaoming Liu

Robotic Intelligence

🎯 What it does: Developed and manufactured a micro piezoelectric robot named DMPBot with a carbon fiber base, capable of achieving high speed, high precision, and omnidirectional motion.

DNAct: Diffusion Guided Multi-Task 3D Policy Learning

Ge Yan, Xiaolong Wang

Robotic IntelligenceReinforcement LearningVision-Language-Action ModelDiffusion modelNeural Radiance Field

🎯 What it does: Propose the DNAct framework, combining neural rendering pre-training with diffusion training to achieve multi-task 3D policy learning through language conditioning.

DnD Filter: Differentiable State Estimation for Dynamic Systems using Diffusion Models

Ziyu Wan, Lin Zhao

Autonomous DrivingDiffusion modelImage

🎯 What it does: Proposed the DnD Filter, a differentiable filter that uses diffusion models for dynamic system state estimation.

Do Visual-Language Grid Maps Capture Latent Semantics?

Matti Pekkanen, Ville Kyrki

Vision Language ModelPoint Cloud

🎯 What it does: This paper proposes a method for evaluating the quality of maps constructed using vision-language model (VLM) embeddings, focusing on two key attributes: map queryability and distinctness, along with corresponding metrics.

DogLegs: Robust Proprioceptive State Estimation for Legged Robots Using Multiple Leg-Mounted IMUs

Yibin Wu, H. Kuhlmann

Robotic Intelligence

🎯 What it does: Propose a state estimation system named DogLegs, which achieves robust body pose estimation for legged robots by fusing body IMU, joint encoders, and multi-leg IMUs.

Dom, cars don’t fly!—Or do they? In-Air Vehicle Maneuver for High-Speed Off-Road Navigation

Anuj Pokhrel, Xuesu Xiao

Autonomous DrivingPhysics Related

🎯 What it does: Studied methods for vehicle aerial maneuvering in high-speed off-road navigation, proposing a fixed-time domain sampling motion planner based on a hybrid forward dynamics model to achieve accurate landing posture within a short time.

Domain-Conditioned Scene Graphs for State-Grounded Task Planning

Jonas Herzog, Yue Wang

Robotic IntelligenceVision Language ModelMultimodality

🎯 What it does: This paper proposes and implements a domain-conditioned scene graph-based state induction framework for perception state induction and subsequent state-based planning in robot task planning; meanwhile, it provides a lightweight vision-language implementation scheme that classifies domain-specific predicates based on domain-related object detection to generate scene graphs; through this framework, scene graphs can be directly mapped to symbolic states in planning languages (e.g., PDDL).

DPGLA: Bridging the Gap between Synthetic and Real Data for Unsupervised Domain Adaptation in 3D LiDAR Semantic Segmentation

Wanmeng Li, Alberto Pretto

SegmentationDomain AdaptationPoint Cloud

🎯 What it does: Proposes a dynamic pseudo-label filtering (DPLF) scheme, a prior-guided data augmentation pipeline (PG-DAP), and a data-mixing consistency loss to enhance the performance of unsupervised domain adaptation for 3D LiDAR point cloud semantic segmentation.

DPGP: A Hybrid 2D-3D Dual Path Potential Ghost Probe Zone Prediction Framework for Safe Autonomous Driving

Weiming Qu, D. Luo

Depth EstimationAutonomous DrivingImage

🎯 What it does: Propose the DPGP framework, which predicts ghost detection regions through 2D-3D fusion using a monocular camera.

DPR-Splat: Depth and Pose Refinement with Sparse-View 3D Gaussian Splatting for Novel View Synthesis

Lingxiang Hu, Ran Song

GenerationPose EstimationDepth EstimationGaussian Splatting

🎯 What it does: Propose the DPR-Splat framework, which utilizes a sparse-view 3D Gaussian projection model and improves the accuracy of camera poses and depth maps through pose and depth refinement modules, thereby achieving higher quality view synthesis.

DPSN: Dual Prior Knowledge Induced Tactile paving and Obstacle Joint Segmentation Network

Youqi Song, Gaoqi He

SegmentationComputational EfficiencyImage

🎯 What it does: Propose a dual prior knowledge induced tactile paving and obstacle joint segmentation network (DPSN), achieving the integration of complete tactile paving masks and obstacle masks, enabling precise segmentation of tactile paving and obstacles.

DR-MPC: Disturbance-Resilient Model Predictive Visual Servoing Control for Quadrotor UAV Pipeline Inspect

Wen Li, Shihua Li

OptimizationRobotic IntelligenceImage

🎯 What it does: Proposed and implemented a pipeline inspection method for quadrotor drones based on disturbance-recoverable model predictive control (DR-MPC), integrating nonlinear MPC with image-based visual servoing, and introducing a generalized extended state observer to estimate disturbances, verifying its effectiveness in the Gazebo simulation environment.

DRACo-SLAM2: Distributed Robust Acoustic Communication-efficient SLAM for Imaging Sonar Equipped Underwater Robot Teams with Object Graph Matching

Yewei Huang, Brendan Englot

Robotic IntelligenceSimultaneous Localization and MappingImage

🎯 What it does: Propose DRACo-SLAM2, a distributed SLAM framework for underwater robot teams equipped with multi-beam imaging sonar, improving the original DRACo-SLAM by adopting object graph representation of sonar maps and object graph matching to achieve fast inter-robot loop closure detection.

DRARL: Disengagement-Reason-Augmented Reinforcement Learning for Efficient Improvement of Autonomous Driving Policy

Weitao Zhou, Diange Yang

Autonomous DrivingReinforcement Learning

🎯 What it does: This paper proposes the DRARL method, which improves autonomous driving strategies by identifying the causes of disengagement.

Drive&Gen: Co-Evaluating End-to-End Driving and Video Generation Models

Jiahao Wang, C. Jiang

GenerationData SynthesisAutonomous DrivingVideo

🎯 What it does: This paper proposes the Drive&Gen framework, which utilizes an end-to-end driving model to assess the authenticity of video generation models and enhances the generalization capability of the driving model by generating synthetic data.

DriveBLIP2: Attention-Guided Explanation Generation for Complex Driving Scenarios

Shihong Ling, Na Du

Autonomous DrivingExplainability and InterpretabilityTransformerVision Language ModelVideo

🎯 What it does: Propose the Drive-Blip2 framework, which generates accurate and context-related explanations for complex driving scenarios based on BLIP2-OPT.

DriveGen: Towards Infinite Diverse Traffic Scenarios with Large Models

Shenyu Zhang, Weinan Zhang

GenerationData SynthesisAutonomous DrivingTransformerLarge Language ModelVision Language ModelDiffusion modelRetrieval-Augmented Generation

🎯 What it does: Propose the DriveGen framework, which employs a two-phase process: in the initialization phase, large language models and retrieval techniques are used to generate maps and vehicle assets, while in the rollout phase, vision-language models and a custom diffusion planner output trajectories; simultaneously, the DriveGen-CS automatic corner case generation pipeline is developed.

DriveLMM-o1: A Step-by-Step Reasoning Dataset and Large Multimodal Model for Driving Scenario Understanding

Ayesha Ishaq, Salman H. Khan

Autonomous DrivingLarge Language ModelSupervised Fine-TuningVision Language ModelMultimodalityChain-of-Thought

🎯 What it does: Proposed a step-by-step visual reasoning dataset called DriveLMM-o1 for autonomous driving scenarios, and fine-tuned a large multimodal model on this dataset to provide a step-by-step reasoning process and answers.

DroneKey: Drone 3D Pose Estimation in Image Sequences using Gated Key-representation and Pose-adaptive Learning

Seo-Bin Hwang, Yeong-Jun Cho

Pose EstimationTransformerVideo

🎯 What it does: Propose the DroneKey framework, combining 2D keypoint detection with 3D pose estimation to address the challenge of drone keypoint detection.

DRP: A Decomposition-Reflection-Prediction Framework for Long-Horizon Robot Task Planning using Large Language Models

Zhaowen Zheng, Jing Wang

Robotic IntelligenceTransformerLarge Language ModelPrompt EngineeringText

🎯 What it does: Proposes a robot task planning framework DRP based on large language models (LLMs), supporting environment knowledge injection to address feasibility issues in long-term tasks.

DRTT : A Diffusion-based Framework for 4DCT Generation, Robust Thoracic Registration and Tumor Deformation Tracking

Dongyuan Li, Xiaojun Chen

Object TrackingGenerationDiffusion modelPoint CloudBiomedical DataComputed Tomography

🎯 What it does: Proposes a framework based on the Recursive Deformable Diffusion Model (RDDM) for real-time tumor tracking, registration, and navigation during surgery, reducing training complexity and improving the utilization of 4D-CT data;

DSFormer-RTP: Dynamic-stream Transformers for Real-time Deterministic Trajectory Prediction

Xun Chen, Danwei Wang

Autonomous DrivingComputational EfficiencyTransformerTime SeriesSequential

🎯 What it does: Proposed a dynamic flow Transformer architecture for real-time deterministic trajectory prediction, treating the prediction task as a sequence-to-sequence model with a single precise output.

DTactive: A Vision-Based Tactile Sensor with Active Surface

Jikai Xu, Huazhe Xu

Robotic IntelligenceImage

🎯 What it does: Designed and implemented the DTactive visual tactile sensor with an active surface, capable of simultaneously performing tactile perception and in-hand object rolling operations, and utilizing learning methods to achieve precise angular trajectory control.

Dual-Arm Hierarchical Planning for Laboratory Automation: Vibratory Sieve Shaker Operations

Haoran Xiao, Wei Dai

OptimizationRobotic Intelligence

🎯 What it does: This paper addresses three key tasks in the operation of vibration screening machines in material laboratories: bimanual can lid operations, bimanual handover in overlapping workspaces, and obstructed powder sample container delivery. A hierarchical planning framework is proposed to solve issues such as low sampling efficiency in narrow channels, insufficient trajectory smoothness, and suboptimal paths.

Dual-Arm Teleoperated Robotic Microsurgery System with Live Volumetric OCT Image Feedback

Jiawei Liu, Mark Draelos

Robotic IntelligenceBiomedical Data

🎯 What it does: A dual-arm remote-controlled minimally invasive surgical robot system equipped with real-time volumetric optical coherence tomography (OCT) image feedback for micro-scale tissue manipulation

Dual-Bubble Coordinated Acoustic Micromanipulator for Multidirectional Object Rotation*

Yuyang Li, Xiaoming Liu

Robotic IntelligenceUltrasound

🎯 What it does: Developed a dual-bubble acoustic micromanipulator capable of achieving multi-directional rotation of microscopic objects.

Dual-Level Open-Vocabulary 3D Scene Representation for Instance-Aware Robot Navigation

Tianlu Zheng, Qichuan Ding

Robotic IntelligenceLarge Language ModelVision Language Model

🎯 What it does: Propose the Dual-Level Open-Vocabulary 3D (DLOV-3D) framework, which integrates pixel-level and image-level features to construct spatial scene representations, enabling instance-aware robot navigation based on free-form queries, and supports integration with LLMs for long-sequence multi-instance navigation.

Dual-Modal Magnetic Skin for Robust Tactile Sensing

Pengwen Xiong, P. X. Liu

Robotic IntelligenceConvolutional Neural NetworkMultimodality

🎯 What it does: Propose a dual-mode soft magnetic skin capable of simultaneously capturing magnetic and mechanical tactile information, and achieve fusion through a CNN-CNN-MLP architecture.

Dual-Mode Motion Control of Multi-Stimulus Deformable Miniature Robots with Adaptive Orientation Compensation in Unstructured Environments

Shihao Zhong, Huaping Wang

Robotic Intelligence

🎯 What it does: Developed a multi-stimuli responsive, deformable microrobot and proposed an adaptive multimodal motion control method.

Dual-Mode Passive Fault-Tolerant Control for Underwater Vehicles with Actuator Faults and Time-Varying Disturbances

Yizong Chen, Yaonan Wang

OptimizationRobotic Intelligence

🎯 What it does: A dual-mode passive fault-tolerant control scheme is proposed for underwater vehicles affected by time-varying external disturbances and actuator faults.

DualAD: Dual-Layer Planning for Reasoning in Autonomous Driving

Dingrui Wang, Johannes Betz

Autonomous DrivingTransformerLarge Language ModelText

🎯 What it does: Proposes the DualAD framework, which simulates human driving reasoning and comprises a lower-level rule-based motion planner and an upper-level text encoder combined with an LLM decision layer.

DualCLIP: Bridging 3D Geometry and Multimodal Semantics for Robotic Perception

Yinghao Liu, Nieqing Cao

Robotic IntelligenceTransformerVision Language ModelContrastive LearningMultimodality

🎯 What it does: Propose the DualCLIP framework, integrating depth-aligned CLIP encoders to achieve collaborative learning between 3D geometry and multimodal semantics;

DuLoc: Life-Long Dual-Layer Localization in Changing and Dynamic Expansive Scenarios

Haoxuan Jiang, Jun Ma

Autonomous DrivingSimultaneous Localization and MappingPoint Cloud

🎯 What it does: Proposes DuLoc, a two-layer localization framework that tightly couples LiDAR-inertial odometry with offline map localization and introduces a constant velocity motion model to suppress noise, enabling seamless integration of prior global maps with real-time local maps for robust localization in unbounded and dynamically changing environments.

DW-VIO: Deep Weighted Visual-Inertial Odometry

Guyuan Chen, Zhaopeng Cui

Pose EstimationOptimizationSimultaneous Localization and MappingMultimodality

🎯 What it does: Developed a learning-based visual inertial odometry system, DW-VIO, capable of fusing multi-sensor data and providing robust state estimation.

Dyna-LfLH: Learning Agile Navigation in Dynamic Environments from Learned Hallucination

Saad Abdul Ghani, Xuesu Xiao

Autonomous DrivingRobotic Intelligence

🎯 What it does: Propose a self-supervised method, Dyna-LfLH, for training motion planners to achieve agile navigation in dense and dynamic obstacle environments.

Dynamic Action Localization and Recognition for Intelligent Perception of Surgical Robots

Yaqin Peng, Qiang Ye

RecognitionRobotic IntelligenceRecurrent Neural NetworkGraph Neural NetworkContrastive LearningBiomedical Data

🎯 What it does: Proposes a self-supervised learning-based method for surgical action recognition, utilizing dynamic masking and attention mechanisms to locate key action regions, and employing graph-enhanced adaptive feature selection to capture temporal relationships; meanwhile, combining LSTM long-term dependency modeling with multi-view contrastive learning to support dynamic adjustment of surgical perspectives and visual navigation;

Dynamic Layer Detection of Thin Materials using DenseTact Optical Tactile Sensors

A. Dhawan, Monroe Kennedy

ClassificationRobotic IntelligenceTransformerOptical FlowMultimodalityTime Series

🎯 What it does: Propose a post-grasping dynamic wiping method utilizing the DenseTact 2.0 optical tactile sensor to classify the number of layers of grasped thin materials.

Dynamic Modeling and Efficient Data-Driven Optimal Control for Micro Autonomous Surface Vehicles

Zhiheng Chen, Wei Wang

Autonomous DrivingOptimizationPhysics Related

🎯 What it does: Proposed a physics-driven dynamic model for micro autonomous surface vehicles (MicroASV) and developed a data-driven optimal control framework based on weak-form online model learning.

Dynamic Network Topology Analysis, Design, and Evaluation for Multi-Robot Vehicle Transfer in High-Density Storage Yards

Lin Zhang, Junzheng Wang

OptimizationRobotic IntelligenceGraph

🎯 What it does: A dynamic network topology framework is proposed to optimize large-scale vehicle transfers in high-density storage areas, utilizing an event-triggered mechanism to update the network topology in real-time, enhancing the routing flexibility and efficiency of robot scheduling.