arXivSub Start free trial

ICRA 2023 Papers with AI Summaries

IEEE International Conference on Robotics and Automation · 1341 papers

3-D Reconstruction Using Monocular Camera and Lights: Multi-View Photometric Stereo for Non-Stationary Robots

Monika Roznere, Alberto Quattrini Li

Depth EstimationOptimizationRobotic IntelligenceSimultaneous Localization and MappingImage

🎯 What it does: Proposes an underwater multi-view photometric stereo (MVPS) framework using a monocular camera and fixed lighting on a non-static low-cost robot for 3D reconstruction.

3-Dimensional Sonic Phase-invariant Echo Localization

Christopher Hahne

Pose EstimationDepth EstimationContrastive LearningAudio

🎯 What it does: Proposed and implemented the Parallax among Corresponding Echoes (PaCE) method for 3D triangulation of acoustic time-of-flight (ToF) pulses in 3D space from arbitrary sensor positions, achieving phase-invariant 3D object localization with only one isotropic transmitter and at least three ToF receivers, with relatively loose sensor position constraints.

3D Reconstruction of Tibia and Fibula using One General Model and Two X-ray Images

Kai Pan, Qi Luo

GenerationOptimizationBiomedical Data

🎯 What it does: Reconstruct patient-specific 3D models of the tibia and fibula from two X-ray images and a generic 3D model

3D Reconstruction-Based Seed Counting of Sorghum Panicles for Agricultural Inspection

Harry Freeman, G. Kantor

ImagePoint CloudMeshAgriculture Related

🎯 What it does: A method for rice panicle seed counting based on 3D reconstruction is proposed, utilizing seeds as 2D and 3D semantic landmark points to generate high-quality 3D models, and estimating the total number of seeds and weight within the panicle through multi-view 2D counting fusion.

3D Spectral Domain Registration-Based Visual Servoing

Maxime Adjigble, Naresh Marturi

OptimizationRobotic IntelligencePoint Cloud

🎯 What it does: Propose a 3D point cloud visual servoing scheme based on spectral domain registration, utilizing 3D FFT to estimate translation, real spherical harmonics to estimate rotation, achieving a decoupled 6-DoF controller, and realizing robotic arm positioning tasks through gradient ascent optimization.

3D VSG: Long-term Semantic Scene Change Prediction through 3D Variable Scene Graphs

Sam Looper, L. Schmid

Robotic IntelligenceGraph Neural NetworkSupervised Fine-TuningPoint Cloud

🎯 What it does: This paper proposes the Variable Scene Graph (VSG) model and its variability estimation method DeltaVSG for predicting long-term semantic changes in indoor scenes.

3D-DAT: 3D-Dataset Annotation Toolkit for Robotic Vision

M. Suchi, M. Vincze

Robotic IntelligenceNeural Radiance FieldImage

🎯 What it does: Proposed a 3D dataset annotation tool called 3D-DAT based on pure RGB, utilizing Neural Radiance Fields (NeRF) to achieve scene-level automatic object alignment and annotation;

3D-Printed Adaptive Microgripper Driven by Thin-Film NiTi Actuators

Sukjun Kim, S. Bergbreiter

Robotic Intelligence

🎯 What it does: Designed and manufactured a 3D printed adaptive micro gripper driven by thin film NiTi micro-actuators, and demonstrated micro-scale shape sorting by integrating it onto a PCB.

3DMODT: Attention-Guided Affinities for Joint Detection & Tracking in 3D Point Clouds

Jyoti Kini, M. Shah

Object DetectionObject TrackingTransformerPoint Cloud

🎯 What it does: Proposed a single-shot end-to-end trainable network for joint multi-object detection and tracking in 3D point clouds, improving the similarity matrix through attention mechanisms.

3DSGrasp: 3D Shape-Completion for Robotic Grasp

S. S. Mohammadi, J. Santos-Victor

Robotic IntelligenceTransformerPoint Cloud

🎯 What it does: Propose the 3DSGrasp strategy, which uses a Transformer encoder-decoder network to predict missing geometry, achieving a complete 3D point cloud and thereby generating reliable grasping poses.

4DRadarSLAM: A 4D Imaging Radar SLAM System for Large-scale Environments based on Pose Graph Optimization

Jun Zhang, Danwei W. Wang

Autonomous DrivingOptimizationSimultaneous Localization and MappingPoint Cloud

🎯 What it does: Proposed a 4D radar-based SLAM system that realizes a complete loop closure process from radar point clouds to pose graphs.

6D Pose Estimation for Textureless Objects on RGB Frames using Multi-View Optimization

Jun Yang, Steven L. Waslander

Pose EstimationOptimizationImage

🎯 What it does: Propose a framework that decomposes 6D pose estimation into first estimating 3D translation and then estimating 3D rotation, leveraging multi-view RGB images.

A Benchmark for Multi-Robot Planning in Realistic, Complex and Cluttered Environments

Simon Schaefer, A. Kleiner

Robotic IntelligenceBenchmark

🎯 What it does: Proposes an open-source benchmark suite for multi-robot planning and coordination in realistic, complex, and crowded 3D simulation environments.

A Bioinspired Synthetic Nervous System Controller for Pick-and-Place Manipulation

Yanjun Li, R. Quinn

Explainability and InterpretabilityRobotic Intelligence

🎯 What it does: This paper uses a learning algorithm to construct compact sub-networks for implementing addition, subtraction, multiplication, and division, and combines them with a bio-inspired Synthetic Nervous System (SNS) to design an interpretable pick-and-place controller. The controller was successfully implemented on a simulated gantry system and migrated to a real robot platform without further parameter tuning.

A causal decoupling approach to efficient planning for logistics problems with stateful stochastic demand

Dipta Chaudhuri, Dylan A. Shell

OptimizationTime Series

🎯 What it does: Propose a method based on causal disentanglement to reduce the planning space in logistics planning with time-related stochastic demand, providing a series of solutions ranging from fully preserving time correlation to partially weakening time correlation; experimental validation was conducted on small-scale cases in lean manufacturing and product routing.

A Coarse-to-Fine Framework for Dual-Arm Manipulation of Deformable Linear Objects with Whole-Body Obstacle Avoidance

Mingrui Yu, Xiang Li

Robotic Intelligence

🎯 What it does: Propose a coarse-to-fine framework that combines global planning with local control, enabling a dual-arm manipulator to operate deformable linear objects in cluttered environments with obstacles, achieving the desired shape while avoiding collisions.

A Compact, Two-Part Torsion Spring Architecture

Zachary Bons, Elliott J. Rouse

Physics Related

🎯 What it does: Proposed and tested a lightweight, compact, and customizable dual-rotational spring design, and experimentally verified its performance

A Complete Set of Connectivity-aware Local Topology Manipulation Operations for Robot Swarms

Koresh Khateri, G. Beltrame

Robotic IntelligenceGraph

🎯 What it does: Demonstrated a complete set of local topological control operations that can transform any initial spanning tree into any connected tree with the same number of nodes, and provided operational sequences to construct linear or star-shaped trees from any spanning tree.

A congestion-aware path planning method considering crowd spatial-temporal anomalies for long-term autonomy of mobile robots

Zijian Ge, M. Coombes

OptimizationRobotic Intelligence

🎯 What it does: Proposed a congestion-aware path planning method that utilizes known spatiotemporal crowd patterns and a partial update memory (PUM) model to generate multi-layer congestion density maps, enabling robots to navigate through less crowded areas to reach their destinations.

A Consistency-Based Loss for Deep Odometry Through Uncertainty Propagation

H. Damirchi, B. Moshiri

Pose EstimationSimultaneous Localization and MappingImage

🎯 What it does: Proposes a loss function for depth odometry networks based on consistency and weighted through uncertainty propagation, allowing the network to adaptively balance incremental and cumulative errors within a maximum likelihood framework;

A Contextual Bandit Approach for Learning to Plan in Environments with Probabilistic Goal Configurations

S. Rudra, G. Aggarwal

OptimizationRobotic IntelligenceReinforcement Learning

🎯 What it does: Propose a modular framework capable of efficiently searching for static and movable target objects in indoor environments; use a Contextual-Bandit agent for exploration, learning the probability of observing targets at different navigable locations, and input these probabilities as rewards into a weighted minimum latency solver to derive the robot's trajectory.

A Continuous Off-Policy Reinforcement Learning Scheme for Optimal Motion Planning in Simply-Connected Workspaces

Panagiotis Rousseas, K. Kyriakopoulos

OptimizationRobotic IntelligenceReinforcement Learning

🎯 What it does: In a simply connected workspace, an offline iterative method is proposed using the Integral Reinforcement Learning framework, which provides safe, convergent, and almost globally optimal motion planning strategies.

A Control Approach for Human-Robot Ergonomic Payload Lifting

Lorenzo Rapetti, D. Pucci

Robotic IntelligenceTime Series

🎯 What it does: Proposed a collaborative robot control method based on full-body wearable sensor monitoring and coupled rigid body dynamics modeling, achieving online trajectory advancement adaptation for human-robot collaboration load enhancement.

A Curvature and Trajectory Optimization-based 3D Surface Reconstruction Pipeline for Ultrasound Trajectory Generation

Ananya Bal, H. Choset

OptimizationPoint CloudBiomedical DataUltrasound

🎯 What it does: Proposed a complete workflow based on RGB-D sensors for 3D surface reconstruction and ultrasound scanning trajectory generation, used for automatically locating the scanning area of the femoral artery in the lower limb.

A Decoupled and Linear Framework for Global Outlier Rejection over Planar Pose Graph

Tianyue Wu, Fei Gao

OptimizationGraph

🎯 What it does: Propose a robust framework for handling loop closure anomalies in 2D pose graph optimization (PGO). The framework first decouples the robust PGO problem wrapped with a truncated least squares kernel into two subproblems, then linearizes the original subproblem represented by rotation matrices through linear angular representation, and subsequently solves the two non-convex subproblems sequentially using the Graduated Non-Convexity (GNC) algorithm without requiring initial guesses.

A Deep Learning Human Activity Recognition Framework for Socially Assistive Robots to Support Reablement of Older Adults

Fraser Robinson, G. Nejat

ClassificationRecognitionRobotic IntelligenceMultimodality

🎯 What it does: This paper proposes a deep learning framework that utilizes multi-modal inputs to achieve autonomous recognition and classification of activities of daily living (ADLs) in elderly people's home environments, aiming to support the long-term deployment of social assistant robots and help the elderly maintain independent living.

A Digital Twin for Teleoperation of Vehicles in Urban Environments

Philipp Kremer, Sangyoung Park

Autonomous DrivingWorld ModelVideo

🎯 What it does: Proposed and implemented a real-time digital twin concept for remote operation of vehicles in urban environments, validated on a vehicle simulator and remote control hardware setup to reduce video transmission volume and reconstruct scenes;

A Distributed Online Optimization Strategy for Cooperative Robotic Surveillance

Lorenzo Pichierri, G. Notarstefano

OptimizationRobotic Intelligence

🎯 What it does: Proposed a distributed online optimization strategy to control a team of collaborative robots for protecting a target from attackers.

A Dual-Arm Participated Human-Robot Collaboration Method for Upper Limb Rehabilitation of Hemiplegic Patients

Lufeng Chen, Hong Cheng

Pose EstimationRobotic Intelligence

🎯 What it does: Propose a new dual-arm collaboration framework that integrates healthy/affected limbs with robots to achieve personalized upper limb rehabilitation training for stroke patients.

A Fast Geometric Framework for Dynamic Cosserat Rods with Discrete Actuated Joints

Hossain Samei, R. Chhabra

Computational EfficiencyPhysics RelatedOrdinary Differential Equation

🎯 What it does: Developed a fast geometric framework within the geometric framework, using finite difference methods to solve the discrete actuator joint model of dynamic Cosserat rods, and conducted numerical experiments and case studies.

A fast two-stage approach for multi-goal path planning in a fruit tree

Werner Kroneman, A.F. van der Stappen

OptimizationAgriculture Related

🎯 What it does: Propose a fast two-stage method for multi-target path planning in orchards, enabling drones equipped with robotic arms to access over 150 targets;

A Flexible 3D Force Sensor with In-Situ Tunable Sensitivity

James Davies, Thanh Nho Do

Robotic Intelligence

🎯 What it does: Developed a flexible 3D force sensor with adjustable sensitivity in vivo.

A fluidic actuator with an internal stiffening structure inspired by mammalian erectile tissue

J. Fras, K. Althoefer

Robotic Intelligence

🎯 What it does: Designed a fluid actuator with an internal hardening structure that utilizes small chambers to capture fluid and achieve localized pressure amplification for directional hardening.

A Force-Sensitive Exoskeleton for Teleoperation: An Application in Elderly Care Robotics

Alexander Toedtheide, Sami Haddadin

Robotic Intelligence

🎯 What it does: Proposed two force-sensitive upper limb exoskeletons for remote operation, demonstrating human-machine interaction control and bidirectional force feedback remote operation of semi-active exoskeletons, as well as designing a three-degree-of-freedom shoulder structure and contact torque elimination controller for future fully active dual-arm exoskeletons;

A Framework for Active Haptic Guidance Using Robotic Haptic Proxies

Niall L. Williams, Ming C Lin

Robotic Intelligence

🎯 What it does: Proposes an active tactile guidance framework using robotic tactile agents, aiming to influence user behavior in mixed reality by actively rendering tactile forces to provide a safer and more immersive experience.

A Framework for Fast Prototyping of Photo-realistic Environments with Multiple Pedestrians

Sara Casao, E. Montijano

GenerationData SynthesisImage

🎯 What it does: Developed a photo-realistic framework based on Unreal Engine and AirSim for rapidly constructing multi-pedestrian, multi-mobile robot scenarios, capable of generating random or customized pedestrian trajectories, and providing 50 human models along with metadata APIs.

A Framework for Simultaneous Workpiece Registration in Robotic Machining Applications

S. Lloyd, M. Ahmadi

Robotic IntelligenceSimultaneous Localization and Mapping

🎯 What it does: Proposed the SRAM framework, which utilizes real-time data to achieve parallel correction of workpiece registration and tool paths, and adaptively adjusts the force controller based on real-time covariance estimation.

A Framework for the Unsupervised Inference of Relations Between Sensed Object Spatial Distributions and Robot Behaviors

Christopher Morse, Sebastian G. Elbaum

Representation LearningRobotic IntelligenceSequential

🎯 What it does: Proposes the SpRinG framework, which uses system trajectories for unsupervised inference to uncover the relationship between the spatial distribution of perceived objects and the behavior of mobile robots.

A Gaze-Speech System in Mixed Reality for Human-Robot Interaction

J. Prada, Cheol Song

Robotic IntelligenceVision-Language-Action ModelImageMultimodalityAudio

🎯 What it does: Proposed and implemented a gaze-speech interaction system in a mixed reality environment, verifying its time performance improvement in pick-and-place tasks.

A general class of combinatorial filters that can be minimized efficiently

Yulin Zhang, Dylan A. Shell

Optimization

🎯 What it does: Analyze the NP-hard nature of combination filter minimization, identifying two independent factors contributing to the difficulty—the structure of the compatibility relation graph and deterministic constraints—and propose a new constraint repair algorithm that covers and unifies three previously known polynomial-time solvable special subclasses, while expanding new families of filters that can be efficiently reduced.

A General Locomotion Approach for a Novel Multi-legged Spherical Robot

Dun Yang, Yang Yu

Robotic Intelligence

🎯 What it does: Designed a 12-legged radial skeleton spherical robot, proposed a high-precision gait generation algorithm applicable to any multi-legged radial skeleton robots, first proposed a closed-loop control scheme for this type of robot, established a dynamic model considering contact friction, and verified through prototype experiments.

A generic diffusion-based approach for 3D human pose prediction in the wild

Saeed Saadatnejad, Alexandre Alahi

Pose EstimationDiffusion modelSequential

🎯 What it does: Proposes a 3D human pose prediction method based on diffusion models, utilizing a denoising framework to handle noisy inputs and missing elements, applicable for long-term prediction in real-world scenarios;

A generic power wheelchair lumped model in the sagittal plane: towards realistic self-motion perception in a virtual reality simulator

F. Grzeskowiak, S. Guégan

Robotic IntelligenceMultimodality

🎯 What it does: Propose a general kinematic and dynamic model for electric wheelchairs (geometric model + equivalent model), and provide identification methods for n-wheel type wheelchairs. Subsequently, implement and validate a motion cueing algorithm based on this model in a multi-sensory VR driving simulator.

A Graph-Based Optimization Framework for Hand-Eye Calibration for Multi-Camera Setups

D. Evangelista, Alberto Pretto

OptimizationRobotic IntelligenceSimultaneous Localization and Mapping

🎯 What it does: Propose a hand-eye calibration method based on graph optimization, applicable to single-camera and multi-camera robotic systems.

A Gravity Compensation Strategy for On-ground Validation of Orbital Manipulators

M. Stefano, C. Ott

OptimizationRobotic IntelligencePhysics Related

🎯 What it does: Proposed and implemented an optimal strategy for gravity compensation on ground-based orbital manipulators, utilizing optimization to minimize joint torque and calculate the compensating force that the external carrier needs to track, achieving full gravity compensation.

A Handheld Hydraulic Cardiac Catheter with Omnidirectional Manipulator and Touch Sensing

Chi Cong Nguyen, Thanh Nho Do

Robotic IntelligenceBiomedical Data

🎯 What it does: Developed a handheld soft robotic catheter for atrial fibrillation ablation

A Hierarchical Decoupling Approach for Fast Temporal Logic Motion Planning

Ziyang Chen, Zheng Kan

Autonomous DrivingOptimization

🎯 What it does: Proposed and implemented a hierarchical separation framework for rapid motion planning under time logic constraints; the framework divides task planning, decoupling layer, and low-level path planning into three layers and realizes independent operation.

A Hybrid Cable-Driven Robot for Non-Destructive Leafy Plant Monitoring and Mass Estimation using Structure from Motion

Gerry Chen, F. Dellaert

Robotic IntelligenceImageAgriculture Related

🎯 What it does: Designed and experimentally verified a hybrid cable-driven robot equipped with a 4-degree-of-freedom (4-DOF) robotic arm and a camera for non-destructive monitoring and quality estimation of leafy vegetables in vertical aquaponic farms.

A Hybrid Quadratic Programming Framework for Real-Time Embedded Safety-Critical Control

Ryan M. Bena, Quan Nguyen

OptimizationRobotic Intelligence

🎯 What it does: Propose a framework that utilizes hybrid computing to achieve real-time embedded safety-critical control, digitally implementing stable control algorithms and using analog resistor arrays to solve safety-critical quadratic programming.

A Hybrid Steerable Robot with Magnetic Wrist for Minimally Invasive Epilepsy Surgery

Changyan He, E. Diller

Robotic Intelligence

🎯 What it does: A hybrid controllable robot integrating a coaxial tube robot and a magnetic wrist gripper was developed for minimally invasive epilepsy surgery, with its feasibility validated through laboratory and brain phantom tests.

A hydraulic soft robotic detrusor based on an origami design

S. Onorati, A. Menciassi

Robotic IntelligenceBiomedical Data

🎯 What it does: Designed and tested an origami-based hydraulic soft bladder detrusor capable of surrounding the bladder and achieving contraction through water absorption to assist in urination

A Joint Modeling of Vision-Language-Action for Target-oriented Grasping in Clutter

Kechun Xu, R. Xiong

Robotic IntelligenceVision-Language-Action ModelMultimodality

🎯 What it does: Proposed a joint modeling framework for vision, language, and actions for target-oriented grasping tasks.

A Kinematically Redundant (6+1)-dof Hybrid Parallel Robot for Delicate Physical Environment and Robot Interaction (pERI)

Jehyeok Kim, C. Gosselin

Robotic Intelligence

🎯 What it does: Proposed and designed a kinematically redundant hybrid parallel robot with 6+1 degrees of freedom, established kinematic models for its two legs and overall structure, analyzed singularities, and subsequently introduced a grasping mechanism driven by redundant degrees of freedom.

A lightweight high-voltage boost circuit for soft-actuated micro-aerial-robots

Zhijian Ren, Yufeng Chen

Robotic Intelligence

🎯 What it does: Developed a lightweight high-voltage boosting circuit capable of stepping up 7.7V DC input to 600V, 400Hz output, used to drive a 120mg soft dielectric elastic actuator, and achieved takeoff in a 158mg miniature aerial robot.

A Linear and Exact Algorithm for Whole-Body Collision Evaluation via Scale Optimization

Qianhao Wang, Fei Gao

OptimizationRobotic Intelligence

🎯 What it does: Proposed a zero-gap whole-body collision evaluation method based on low-dimensional linear programming, capable of analytically solving with linear complexity.

A Little Bit Attention Is All You Need for Person Re-Identification

M. Eisenbach, H. Groß

RetrievalComputational EfficiencyNeural Architecture SearchTransformer

🎯 What it does: Proposed a lightweight attention module and determined its optimal integration position through neural architecture search to minimize inference speed loss while maintaining accuracy.

A method for selecting stumble recovery response in a knee exoskeleton

Maura E. Eveld, M. Goldfarb

Robotic IntelligenceTime SeriesBiomedical Data

🎯 What it does: Developed a fall detection and recovery strategy identification algorithm for knee exoskeletons, and verified its effectiveness in experiments.

A Method to Use Haptic Feedback of Laryngoscope Force Vector for Endotracheal Intubation Training

Haonan Zhou, Thrishantha Nanayakkara

Robotic Intelligence

🎯 What it does: Designed and verified a tracheal intubation training mannequin with tactile feedback, utilizing a silicone tongue embedded with magnets and Hall effect sensors to achieve real-time force vector measurement and feedback.

A Miniaturised Camera-based Multi-Modal Tactile Sensor

K. Althoefer, Peng Qi

Robotic IntelligenceImageMultimodality

🎯 What it does: A miniaturized hemispherical camera-based multimodal tactile sensor was proposed, capable of achieving precise mechanical measurement and tactile perception within a single system. By constructing a smooth silicone hemispherical sensing medium, uniformly distributing markers on its curved surface, utilizing diffused LEDs for enhanced illumination, and employing a spring-based compact force-sensing mechanical structure, the sensor can real-time capture multi-axis force, local force distribution, and contact geometry information, with end-to-end deep learning methods used for data processing.

A minimum swept-volume metric structure for configuration space

Y. Mont-Marin, J. Laumond

Robotic Intelligence

🎯 What it does: Proposed and demonstrated a metric structure based on swept volumes for geodesic computation and path planning in configuration spaces.

A Model-Based Analysis of The Effect of Repeated Unilateral Low Stiffness Perturbations on Human Gait: Toward Robot-Assisted Rehabilitation

Vaughn Chambers, P. Artemiadis

Robotic IntelligenceBiomedical Data

🎯 What it does: Built and extended a neuromuscular gait model to adapt to unilateral low-stiffness (soft) ground, and validated it using experimental human data.

A Moving Target Tracking System of Quadrotors with Visual-Inertial Localization

Ziyue Lin, Wei Wang

Object TrackingOptimizationSimultaneous Localization and MappingImageMultimodality

🎯 What it does: Implemented a vision-based mobile target tracking system that tracks a quadrotor in indoor environments without GNSS using visual-inertial localization.

A Multi-Agent Approach for Adaptive Finger Cooperation in Learning-based In-Hand Manipulation

Lingfeng Tao, Xiaoli Zhang

Robotic IntelligenceReinforcement Learning

🎯 What it does: This paper treats each finger as an independent agent and employs a multi-agent approach to collaboratively perform in-hand manipulation tasks;

A Multi-step Dynamics Modeling Framework For Autonomous Driving In Multiple Environments

Jason Gibson, Evangelos A. Theodorou

Autonomous DrivingRecurrent Neural Network

🎯 What it does: Proposes a multi-step dynamic prediction framework for the dynamics prediction of off-road vehicles

A MySQL Database for the Systematic Configuration Selection of Redundant Manipulators when Path Planning in Confined Spaces

K. Wood, Antonia Tzemanaki

OptimizationRobotic IntelligenceTabular

🎯 What it does: This paper proposes using the MySQL database to systematically select target joint configurations for redundant manipulators in confined space path planning, and generates a complete configuration database to support path planning.

A Negative Imaginary Theory-Based Time-Varying Group Formation Tracking Scheme for Multi-Robot Systems: Applications to Quadcopters

Yu-Hsiang Su, A. Lanzon

Robotic Intelligence

🎯 What it does: A time-varying formation tracking scheme for multi-robot systems based on negative real (NI) theory was developed, employing a two-loop control structure. The nonlinear dynamics of each multi-robot system were converted into a double integrator system via feedback linearization, and an NI-based time-varying formation control protocol was implemented in the outer loop.

A New Efficient Eye Gaze Tracker for Robotic Applications

Chaitanya Bandi, Ulrike Thomas

Robotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Proposed a CNN-based eye gaze tracking method primarily for robot applications.

A New Robust Control Framework for Robot Manipulators without Velocity Measurements: A Modified Dual-loop Control Scheme

H. Park, Jung Hoon Kim

Robotic Intelligence

🎯 What it does: A new framework for computing torque without speed measurement (CTM) is proposed. First, the CTM is achieved based on the Luenberger observer using only position measurements, and it is pointed out that external disturbances simultaneously affect tracking performance and state estimation accuracy. Subsequently, a new dual-loop control scheme is constructed, which can simultaneously improve tracking performance and estimation accuracy, and the selection criteria for corresponding control parameters are provided. Finally, the effectiveness and practicality of this structure are verified through simulation and experiments.

A New Sensation: Digital Strain Sensing for Disturbance Detection In Flapping Wing Micro Aerial Vehicles

Regan Kubicek, S. Bergbreiter

Anomaly Detection

🎯 What it does: Developed and tested a micro-scale digital strain sensor using two-photon polymer 3D printing on flexible Mylar wings, capable of changing logical states upon reaching a specific strain threshold for detecting wind disturbances in flapping-wing aircraft.

A Non-parametric Skill Representation with Soft Null Space Projectors for Fast Generalization

João Silvério, Yanlong Huang

Robotic Intelligence

🎯 What it does: Proposed a non-parametric motion primitive representation incorporating a soft null-space projector for fast and efficient motion generation and adaptation.

A Non-planar Assembly of Modular Tetrahedral-shaped Aerial Robots

Obadah Wali, Eric Feron

Robotic Intelligence

🎯 What it does: Proposed and designed a modular aerial robot named TetraQuad with a tetrahedral geometric structure, demonstrating the scalability of multi-module vertical assembly, enhanced structural stiffness, and reduced vortex interference. A modular control strategy was developed and validated through simulation, with the design ultimately verified in experimental flights.

A Novel Concentric Tube Steerable Drilling Robot for Minimally Invasive Treatment of Spinal Tumors Using Cavity and U-shape Drilling Techniques

Susheela Sharma, F. Alambeigi

Robotic Intelligence

🎯 What it does: Designed, manufactured, and evaluated a new cylindrical coaxial tube controllable drilling robot (CT-SDR) for improving minimally invasive treatment of spinal tumors.

A Novel Platform to Control Biofouling in Pearl Oysters Cultivation

V. Tran, Sai-Kit Yeung

Agriculture Related

🎯 What it does: Designed and implemented an automated pearl clam farming platform, utilizing an air-water displacement mechanism to periodically bring clams to the surface, exposing attached organisms to air and sunlight, thereby eliminating the laborious need for manual cleaning.

A Passivity-based Approach on Relocating High-Frequency Robot Controller to the Edge Cloud

Xiao Chen, Sami Haddadin

Robotic Intelligence

🎯 What it does: Migrate the unified force-impedance controller (UFIC) to the cloud and establish a high-frequency feedback closed-loop in the communication channel through the time-domain passivity approach (TDPA), ensuring the complete system stability of the robot and environment.

A Performance Optimization Strategy Based on Improved NSGA-II for a Flexible Robotic Fish

Ben Lu, Chao Zhou

OptimizationRobotic Intelligence

🎯 What it does: Established a dynamic model incorporating flexible deformation to predict the speed of robotic fish, and employed the improved NSGA-II (NSGA-II-AMS) to optimize the conflicting objectives of speed and power consumption.

A Plug-In Weight-Shifting Module That Adds Emotional Expressiveness to Inanimate Objects in Handheld Interaction

Yohei Noguchi, F. Tanaka

🎯 What it does: A weight displacement module that can be inserted into various objects was developed and tested, achieving three-dimensional weight displacement through controlled one-dimensional translation and two-dimensional rotation, thereby expressing emotions in scenarios such as toy robots, cushions, and containers.

A Preliminary Study of the Effects of Active Recovery Reflexes on Stumble Recovery in a Swing-Assist Knee Prosthesis

Jantzen T. Lee, M. Goldfarb

OptimizationRobotic IntelligenceBiomedical Data

🎯 What it does: This study investigates the effectiveness of a swing-phase fall recovery controller in a swing-assisted knee prosthetic, conducting a pre-experiment on two transfemoral amputees to compare their falling scenarios under the original prosthetic and the prototype device.

A Probabilistic Framework for Visual Localization in Ambiguous Scenes

Fereidoon Zangeneh, P. Jensfelt

Pose EstimationImage

🎯 What it does: Proposed a probabilistic framework for visual localization in blurred scenes that can predict arbitrary-shaped camera pose posterior distributions for a given image.

A Probabilistic Model of Activity Recognition with Loose Clothing

Tiancheng Shen, M. Howard

Recognition

🎯 What it does: Built and validated a probabilistic model-based activity recognition method, explaining why fabric-attached sensors achieve higher accuracy than rigid-attached sensors within short time windows.

A Probabilistic Rotation Representation for Symmetric Shapes With an Efficiently Computable Bingham Loss Function

H. Sato, K. Nishiwaki

Pose EstimationRepresentation LearningPoint Cloud

🎯 What it does: Proposed a fast computable and easy-to-implement Bingham distribution negative log-likelihood (NLL) loss function, and constructed an inference network for predicting symmetric rotations of target objects from point clouds.

A Reachability Tree-Based Algorithm for Robot Task and Motion Planning

Kanghyun Kim, Min Jun Kim

Robotic Intelligence

🎯 What it does: Proposed a robot task and motion planning algorithm based on reachability trees

A real-time dynamic obstacle tracking and mapping system for UAV navigation and collision avoidance with an RGB-D camera

Zhefan Xu, Kenji Shimada

Object TrackingDepth EstimationAutonomous DrivingImage

🎯 What it does: Proposes a real-time dynamic obstacle tracking and mapping system based on RGB-D cameras for quadrotor navigation and collision avoidance.

A Robotic Cooperative Network for Localising a Submarine in Distress: Results From REPMUS21

G. Ferri, A. Tesei

Robotic IntelligenceSimultaneous Localization and MappingAudio

🎯 What it does: Developed a perception layer based on occupancy grids for a multi-robot underwater drone network equipped with passive sonar, achieving position estimation of a submerged submarine emitting a distress signal by fusing azimuth measurements from each robot.

A Sensitivity-Aware Motion Planner (SAMP) to Generate Intrinsically-Robust Trajectories

Simon Wasiela, T. Siméon

Autonomous DrivingOptimization

🎯 What it does: Proposes a global control-aware motion planner SAMP based on state sensitivity, generating collision-free trajectories robust to parameter uncertainties.

A Sequential Quadratic Programming Approach to the Solution of Open-Loop Generalized Nash Equilibria

Edward L. Zhu, F. Borrelli

Optimization

🎯 What it does: Proposed a numerical method for solving the local generalized Nash equilibrium of open general and dynamic games with nonlinear dynamics and constraints.

A Silicone-sponge-based Variable-stiffness Device

Tianqi Yue, J. Rossiter

🎯 What it does: Designed and fabricated a silicone sponge-based variable stiffness device (SVD) that adjusts its stiffness through pressure differential.

A Social Referencing Disambiguation Framework for Domestic Service Robots

Kevin Fan, K. Dautenhahn

Robotic IntelligenceReinforcement Learning from Human Feedback

🎯 What it does: Proposed and experimentally validated a social reference disambiguation framework for home service robots (Fetch), which includes: detecting and analyzing ambiguity, evaluating human attention, disambiguating through human feedback, and learning new objects after clarification.

A Soft Hybrid-Actuated Continuum Robot Based on Dual Origami Structures

J. Tao, Erbao Dong

Robotic Intelligence

🎯 What it does: Proposed and experimentally verified a flexible continuous robot based on a dual origami structure, achieving long extension and large bending through pneumatic and traction winding dual-mode actuation;

A Soft Robot with Three Dimensional Shape Sensing and Contact Recognition Multi-Modal Sensing via Tunable Soft Optical Sensors

Max McCandless, S. Russo

RecognitionRobotic Intelligence

🎯 What it does: Propose a strategy for preparing flexible optical sensors through roughness regulation, enabling a single multimodal sensor to simultaneously perform three-dimensional shape sensing and contact identification, and integrating it into a fully flexible robotic platform;

A Stiffness-Changeable Soft Finger Based on Chain Mail Jamming

Zhengtao Hu, K. Harada

Robotic Intelligence

🎯 What it does: Developed a variable stiffness soft finger based on chain ring blocking, including design, manufacturing, grip force prediction model, and grip strategy, and experimentally verified its adaptive gripping and in-hand manipulation capabilities.

A Study into Understanding User Requirements to Inform the Design of Customizable Robotic Pain Management Devices

Angela Higgins, P. Caleb-Solly

Robotic IntelligenceText

🎯 What it does: Investigated the needs of chronic pain patients for customizable robotic pain management devices through online surveys and interviews

A Tactile Feedback Insertion Strategy for Peg-in-Hole Tasks

Oliver Gibbons, P. Maiolino

Robotic Intelligence

🎯 What it does: Developed a tactile feedback-based insertion strategy for performing peg-in-hole tasks under uncertain conditions, with real-time compensation for the pose changes of the plug inside the gripper.

A Tactile-enabled Hybrid Rigid-Soft Continuum Manipulator for Forceful Enveloping Grasps via Scale Invariant Design

Ian H. Taylor, Alberto Rodriguez

Robotic Intelligence

🎯 What it does: Proposed and implemented a hybrid rigid-soft continuum manipulator integrating rigidity and flexibility with high-resolution tactile perception, achieving strong, compliant, safe, and controllable grasping.

A Task Allocation Framework for Human Multi-Robot Collaborative Settings

M. Lippi, A. Marino

OptimizationRobotic Intelligence

🎯 What it does: Proposes a task allocation framework for human-robot collaborative scenarios, combining offline optimal allocation with online reallocation strategies.

A Trajectory Planner For Mobile Robots Steering Non-Holonomic Wheelchairs In Dynamic Environments

Martin Schulze, R. Dillmann

Robotic Intelligence

🎯 What it does: Propose a trajectory planner for a non-omnidirectional passive wheelchair controlled by a mobile omnidirectional robot;

A Unified BEV Model for Joint Learning of 3D Local Features and Overlap Estimation

Lin Li, Guowei Wan

Autonomous DrivingConvolutional Neural NetworkPoint Cloud

🎯 What it does: Propose a unified bird's-eye view (BEV) model for jointly learning 3D local features and overlap estimation to accomplish point cloud registration and loop closure detection.

A Virtual Reality Framework For Fast Dataset Creation Applied to Cloth Manipulation with Automatic Semantic Labelling

Júlia Borràs Sol, C. Torras

Data SynthesisVideo

🎯 What it does: A human clothing folding demonstration dataset was collected and publicly released using a virtual reality framework, and automatic semantic labeling was achieved through low-dimensional fabric representation.

A Virtual Reality Planning Environment for High-Risk, High-Latency Teleoperation

Will Pryor, P. Kazanzides

Robotic Intelligence

🎯 What it does: Developed a 3D VR interface based on the Meta Quest 2 HMD for the IPSE system, and compared it with the previous 2D keyboard and mouse interface.

AANet: Aggregation and Alignment Network with Semi-hard Positive Sample Mining for Hierarchical Place Recognition

Feng Lu, Chun Yuan

RecognitionContrastive LearningImage

🎯 What it does: Proposed a unified AANet network, combining a global feature aggregation module and a dynamic alignment of local features module to achieve hierarchical visual localization, and introduced a semi-hard positive sample mining strategy to enhance network robustness.

Accelerating Multi-Agent Planning Using Graph Transformers with Bounded Suboptimality

Chenning Yu, Amanda Prorok

OptimizationGraph Neural NetworkTransformerGraph

🎯 What it does: Propose using a Graph Transformer as a heuristic function for Conflict-Based Search (CBS) to accelerate multi-agent path planning while ensuring completeness and bounded suboptimality.