arXivSub Start free trial

IROS 2023 Papers — Page 2

IEEE/RSJ International Conference on Intelligent Robots and Systems · 1195 papers

Agent Prioritization and Virtual Drag Minimization in Dynamical System Modulation For Obstacle Avoidance of Decentralized Swarms

Louis-Nicolas Douce, A. Billard

Robotic IntelligenceAgentic AI

🎯 What it does: Improved the obstacle avoidance algorithm based on dynamic system modulation in heterogeneous omnidirectional mobile agent groups, proposing smooth priority modulation, soft decoupled rotation control, virtual resistance, and designing a safety module to avoid conflicts, followed by verification in simulated assistive living and hospital environments.

Aggregating Single-Wheeled Mobile Robots for Omnidirectional Movements

Meng Wang, Hangxin Liu

OptimizationRobotic Intelligence

🎯 What it does: A self-reconfigurable modular single-wheel mobile robot system was constructed, utilizing steerable omnidirectional wheels and magnetic docking to enable multi-module collaboration to encircle the target object and achieve omnidirectional movement; an optimized wheel steering distribution method was proposed, and hierarchical control was implemented in simulation and experiments, verifying the trajectory tracking performance of single modules and six-module teams in multi-navigation and collaborative transportation tasks.

Aggressive Trajectory Generation for a Swarm of Autonomous Racing Drones

Yu-heng Shen, S. Li

OptimizationRobotic Intelligence

🎯 What it does: Proposed and verified a time-optimal trajectory generation method for quadrotor drone swarms, enabling them to pass through predefined waypoints with maximum maneuverability without collisions.

Air-M: A Visual Reality Many-Agent Reinforcement Learning Platform for Large-Scale Aerial Unmanned System

Jiabin Lou, Rongye Shi

Robotic IntelligenceReinforcement Learning

🎯 What it does: Proposed the Air-M platform, which supports large-scale drone swarm reinforcement learning in a distributed Docker container environment, and bridges virtual reality scenes with the real world.

AirLine: Efficient Learnable Line Detection with Local Edge Voting

Xiao Lin, Chen Wang

Convolutional Neural Network

🎯 What it does: Proposes AirLine, a learnable edge-based line segment detection algorithm that can directly extract line segments from edges and achieve efficient line parameterization through region growth and local edge voting.

AirVO: An Illumination-Robust Point-Line Visual Odometry

Kuan Xu, Lihua Xie

Pose EstimationConvolutional Neural NetworkGraph Neural NetworkSimultaneous Localization and MappingImage

🎯 What it does: Proposed an illumination-robust point-line visual odometry system that integrates accelerated learning-based corner detection (CNN+GNN) and extended line feature matching, with optimized inference processes for real-time operation on low-power embedded platforms.

All Aware Robot Navigation in Human Environments Using Deep Reinforcement Learning

Xiaojun Lu, H. Asama

Robotic IntelligenceReinforcement Learning

🎯 What it does: Propose a fully perceptual neural network that uses deep reinforcement learning to simultaneously handle crowds, obstacles, and individuals, achieving safe and socially compliant navigation for robots in human-populated environments.

AmbiSense: Acoustic Field Based Blindspot-Free Proximity Detection and Bearing Estimation

Siddharth Rupavatharam, Volkan Isler

Safty and PrivacyRobotic IntelligenceAudio

🎯 What it does: Propose the AmbiSense system, which utilizes a single low-cost piezoelectric sensor to create a blind-zone-free acoustic field, enabling proximity detection and direction estimation to enhance human-robot interaction safety.

An Affordances and Electromyography Based Telemanipulation Framework for Control of Robotic Arm-Hand Systems

Ricardo V. Godoy, Minas V. Liarokapis

ClassificationRecognitionObject DetectionRobotic IntelligenceTransformerVision-Language-Action ModelImageMultimodalityBiomedical Data

🎯 What it does: Proposed an EMG remote control framework based on affordances, utilizing an external camera for scene understanding and object detection/identification, providing robotic arm and hand systems with assistance for grasping and manipulation, simplifying user control; meanwhile, adopting object-specific Transformer classifiers to reduce potential gesture outputs, enhancing decoding robustness and accuracy.

An Approach for Generating Families of Energetically Optimal Gaits from Passive Dynamic Walking Gaits

Nelson Rosa, C. Remy

OptimizationRobotic IntelligencePhysics Related

🎯 What it does: Proposed a method for computing a continuous family of locally energy-optimal gaits for bipedal robots using passive gaits.

An Approach to Design a Biomechanically-Inspired Reward Function to Solve a Patience Cube Under Reinforcement Learning Framework

Janghyeon Kim, Hangon Yoon

Robotic IntelligenceReinforcement LearningBiomedical Data

🎯 What it does: Design a reward function from the perspective of control theory and biomechanics, utilizing kinematic and electromyography (EMG) data from human expert demonstrations to train a reinforcement learning (RL) agent to solve the patience blocks problem.

An Attentional Recurrent Neural Network for Occlusion-Aware Proactive Anomaly Detection in Field Robot Navigation

Andre Schreiber, K. Driggs-Campbell

Anomaly DetectionRecurrent Neural NetworkAgriculture Related

🎯 What it does: Proposes an attention-based recurrent neural network for proactive anomaly detection in mobile robots within agricultural environments, integrating current perceptual inputs, planned control actions, and potential representations of previous states.

An Avatar Robot Overlaid with the 3D Human Model of a Remote Operator

Ravi Tejwani, H. Asada

Computational EfficiencyRobotic IntelligenceMesh

🎯 What it does: Developed a mixed reality/virtual Avatar robot system that uses AR to overlay the remote operator's 3D model onto the robot's structure, enabling tactile feedback when the robot is touched.

An Efficient Trajectory Planner for Car-Like Robots on Uneven Terrain

Longji Xu, Fei Gao

OptimizationRobotic Intelligence

🎯 What it does: Proposes a terrain pose mapping and trajectory optimization framework for planning trajectories for vehicle-like robots on uneven terrain.

An Energetic Approach to Task-Invariant Ankle Exoskeleton Control

Katharine Walters, Robert D. Gregg

Robotic IntelligenceBiomedical Data

🎯 What it does: Developed and implemented an ankle exoskeleton controller based on energy shaping, utilizing ankle sensors to achieve task-invariant assistance, and conducted experiments on commercial bilateral ankle exoskeletons; the experimental subjects were three healthy participants walking on a treadmill and in a loop.

An Energy-Efficient Lane-Keeping System Using 3D LiDAR Based on Spiking Neural Network

Genghang Zhuang, A. Knoll

Autonomous DrivingComputational EfficiencySpiking Neural NetworkReinforcement LearningPoint Cloud

🎯 What it does: Propose an end-to-end lane-keeping system based on Spiking Neural Networks (SNN) using 3D LiDAR.

An Ensemble of Online Estimation Methods for One Degree-of-Freedom Models of Unmanned Surface Vehicles: Applied Theory and Preliminary Field Results with Eight Vehicles

Tyler M. Paine, M. Benjamin

Autonomous DrivingOptimizationRecurrent Neural NetworkMixture of ExpertsSequential

🎯 What it does: Experimental evaluation of online system identification for unmanned surface vehicles (USV) was conducted using three mainstream methods and an integrated (ensemble) framework: provably convergent shallow recurrent neural networks (RNN), adaptive identification (AID), and recursive least squares (RLS), with online estimation performed on eight USVs over a total of 30 hours.

An Evaluation of Action Segmentation Algorithms on Bimanual Manipulation Datasets

Andre Meixner, T. Asfour

SegmentationRobotic IntelligenceVideoBenchmark

🎯 What it does: Segment each hand's movements separately to learn a model for bimanual task coordination

An Event-Based Tracking Control Framework for Multirotor Aerial Vehicles Using a Dynamic Vision Sensor and Neuromorphic Hardware

Sotirios N. Aspragkathos, K. Kyriakopoulos

Object TrackingAutonomous DrivingRobotic IntelligenceSpiking Neural Network

🎯 What it does: Proposed an event-based control framework for multirotor drones, utilizing dynamic vision sensors (DVS) and neuromorphic hardware to efficiently track contour regions such as road surfaces.

An Implantable Variable Length Actuator for Modulating in Vivo Musculo-Tendon Force in a Bipedal Animal Model

S. Thomas, Jonas Rubenson

Robotic Intelligence

🎯 What it does: Developed a fully implantable variable-length isometric muscle force actuator to regulate tendon force and replace the lateral gastrocnemius in a bipedal animal model;

An Inflatable Eversible Finger Pad for Variable-Stiffness Grasping with Parallel-Jaw Grippers

Raphael Deimel, Andreas Kugi

Robotic Intelligence

🎯 What it does: Proposed an inflatable and reversible finger pad that allows conventional parallel grippers to change grip stiffness while maintaining contact force and contact with non-planar surfaces.

An Interacting Multiple Model Approach Based on Maximum Correntropy Student's T Filter

Fethi Candan, Lyudmila Mihaylova

Autonomous DrivingRobotic IntelligenceTime Series

🎯 What it does: Proposed an IMM method based on the maximum mutual information Student's T filter for handling non-Gaussian measurement noise, implemented on the Crazyflie 2.0 drone.

An Interactive System for Multiple-Task Linear Temporal Logic Path Planning

Yizhou Chen, Ben M. Chen

Autonomous DrivingRobotic IntelligenceAudio

🎯 What it does: Proposed an interactive system that includes a multi-task linear temporal logic (LTL) path planner and a human-machine interface (HMI). The HMI converts verbal commands into machine-readable task commands, while the planner utilizes Rapidly-exploring Random Trees (RRT) to search for multi-task solutions. During task switching, the planner reinitializes and reconnects the search tree to reuse collected workspace information.

An MCTS-DRL Based Obstacle and Occlusion Avoidance Methodology in Robotic Follow-Ahead Applications

Sahar Leisiazar, Mo Chen

Robotic IntelligenceReinforcement Learning

🎯 What it does: Proposed an obstacle and occlusion avoidance method for robot following applications, and developed a high-level decision algorithm to generate short-term navigation goals.

An Open-Source Robotic Chinese Chess Player

Shan An, Hong Zhang

Object DetectionRobotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Designed and implemented a low-cost, open-source robot for playing Chinese chess.

An Origami-Based Miniature Jumping Robot with Adjustable Jumping Trajectory and Enhanced Intermittent Jumps

Zhipeng Xiong, Bing Li

Robotic Intelligence

🎯 What it does: Proposed a mini jumping robot based on an origami structure, capable of performing intermittent, omnidirectional adjustable jumps, and achieving self-reset and posture adjustment.

An Origami-Inspired Deployable Space Debris Collector

Yuto Tanaka, Ran Dai

🎯 What it does: Designed, actuated, and manufactured a deployable space debris collector based on a conical Kresling origami pattern.

An Orthogonal Collocation Method for Static and Dynamic Cosserat Rods

Radhouane Jilani, Erwan Kerrien

Computational EfficiencyPhysics Related

🎯 What it does: Propose an orthogonal interpolation method (CM) to solve static and dynamic Cosserat rod Dirichlet–Neumann boundary value problems.

Analysis of Lower Extremity Shape Characteristics in Various Walking Situations for the Development of Wearable Robot

Joohyun Park, Hyunki In

Robotic IntelligenceTime Series

🎯 What it does: Measure and analyze changes in thigh and calf circumferences during walking using rope tension sensors and FSR insoles for experiments at multiple speeds and slopes, exploring the impact of these deformations on the design of wearable robot bands;

Analytical Computation of the Contact Force Jacobian for MRI-Actuated Robotic Catheter

Yuttana Itsarachaiyot, M. C. Çavuşoğlu

Robotic IntelligenceMagnetic Resonance ImagingOrdinary Differential Equation

🎯 What it does: Proposed an analytical computation method for the contact force Jacobian matrix of the Cosserat rod model applied to MRI-driven robotic catheters

Analytical Jacobian Approximation for Direct Optimization of a Trajectory of Interpolated Poses on SE(3)

Kazii Botashev, Gonzalo Ferrer

OptimizationPoint Cloud

🎯 What it does: Studied time-continuous trajectory representations based on SE(3) direct linear interpolation, and proposed a new analytical Jacobian approximation method.

ANEC: Adaptive Neural Ensemble Controller for Mitigating Latency Problems in Vision-Based Autonomous Driving

Aws Khalil, Jaerock Kwon

Autonomous DrivingComputational EfficiencyMixture of ExpertsImage

🎯 What it does: Studied the impact of algorithmic latency in vision-driven neural networks for lane-keeping tasks, and proposed the Adaptive Neural Ensemble Controller (ANEC).

Anytime, Anywhere: Human Arm Pose from Smartwatch Data for Ubiquitous Robot Control and Teleoperation

F. Weigend, H. B. Amor

Pose EstimationRobotic IntelligenceTime Series

🎯 What it does: Designed an optimized machine learning method based on a single smartwatch for estimating human arm posture;

AOSoar: Autonomous Orographic Soaring of a Micro Air Vehicle

Sunyou Hwang, G. D. Croon

Autonomous DrivingOptimizationPhysics Related

🎯 What it does: Proposed a fully autonomous mountain gliding method that utilizes an Incremental Nonlinear Dynamic Inversion (INDI) controller with control allocation, and employs simulated annealing optimization to find gliding positions, achieving a maximum flight duration of 30 minutes in wind tunnel experiments with an average throttle usage rate of only 0.25%.

Approximation Algorithms for Charging Station Placement for Mobile Robots

Tanmoy Kundu, Indranil Saha

OptimizationRobotic Intelligence

🎯 What it does: Proves that two categories of charging station location problems are NP-hard, and designs two polynomial-time approximation algorithms to find near-optimal solutions.

Arena-Rosnav 2.0: A Development and Benchmarking Platform for Robot Navigation in Highly Dynamic Environments

Linh Kästner, Jens Lambrecht

Robotic IntelligenceBenchmark

🎯 What it does: Proposed the Arena-Rosnav 2.0 platform, expanding upon previous versions Arena-Bench and Arena-Rosnav by adding multiple modules for development and benchmarking, rearchitecting into a unified API, incorporating more realistic simulations and pedestrian behaviors, and providing detailed documentation; collected user feedback through studies, then integrated two new simulators and various state-of-the-art navigation methods for mutual benchmark comparisons.

ARMP: Autoregressive Motion Planning for Quadruped Locomotion and Navigation in Complex Indoor Environments

Jeonghwan Kim, Sehoon Ha

OptimizationRobotic IntelligenceSupervised Fine-Tuning

🎯 What it does: Proposed the ARMP framework, which can generate arbitrary-length quadruped robot locomotion and navigation action plans in an autoregressive manner, and constructs a large number of trajectory optimization problems to generate an action library. It uses supervised learning to learn the action manifold and integrates it as a low-level controller into the robot navigation framework.

Assignment Algorithms for Multi-Robot Multi-Target Tracking with Sufficient and Limited Sensing Capability

Peihan Li, Lifeng Zhou

Object TrackingOptimizationRobotic Intelligence

🎯 What it does: Propose and analyze a greedy allocation algorithm for single-robot and dual-robot target tracking, optimizing tracking quality and providing approximation guarantees.

Assisting Spectral Mapping Using Cameras

Srinivasan Vijayarangan, David Wettergreen

Robotic IntelligenceImage

🎯 What it does: Utilize RGB cameras to assist in spectral mapping and achieve large-scale detection on robot platforms (especially aerial platforms) to improve the accuracy of spectral reconstruction.

Asynchronous, Option-Based Multi-Agent Policy Gradient: A Conditional Reasoning Approach

Xubo Lyu, Yong Zhang

Reinforcement Learning

🎯 What it does: Proposes an asynchronous, option-based multi-agent policy gradient method based on conditional reasoning to address coordination issues when multiple robots execute options asynchronously.

Attention for Robot Touch: Tactile Saliency Prediction for Robust Sim-to-Real Tactile Control

Yijiong Lin, N. Lepora

Depth EstimationDomain AdaptationRobotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Propose and implement a tactile saliency prediction method based on tactile images, utilizing a three-network system (ConDepNet, TacSalNet, TacNGen) to perform contact depth mapping, saliency mapping, and noise feature generation on high-resolution tactile images, enhancing the robustness of simulated-to-real tactile control in unknown interference environments.

Attention-Based VR Facial Animation with Visual Mouth Camera Guidance for Immersive Telepresence Avatars

Andre Rochow, Sven Behnke

GenerationPose EstimationTransformerImage

🎯 What it does: Propose a hybrid method that combines keypoint detection and oral camera visual guidance to achieve facial animation in VR environments, supporting quick setup and generalization to unseen operators.

Augmentation Enables One-Shot Generalization in Learning from Demonstration for Contact-Rich Manipulation

Xing Li, O. Brock

Robotic Intelligence

🎯 What it does: Proposed a Learning from Demonstrations (LID) method for contact-rich operations, achieving generalization across different mechanisms and environments through autonomous augmentation of a single demonstration.

Augmented Avatar Toward Both Remote Communication and Manipulation Tasks

Masaki Haruna, Susumu Morita

Robotic IntelligenceBenchmark

🎯 What it does: Designed and built an 'enhanced avatar' system capable of simultaneous remote communication and operational tasks, with prototype construction and operational testing conducted.

Augmented Reality Navigation in Robot-Assisted Surgery with a Teleoperated Robotic Endoscope

V. Penza, L. Mattos

Robotic IntelligenceImageMeshBiomedical Data

🎯 What it does: Designed and evaluated an integrated system for real-time AR navigation in robotic minimally invasive surgery, incorporating a robotic endoscope camera, software-based teleoperation achieving remote center of motion (RCM), and AR navigation software based on manual registration.

Autocomplete of 3D Motions for UAV Teleoperation

Batool Ibrahim, Daniel C. Asmar

Recognition

🎯 What it does: Proposes an autocompletion framework for 3D UAV teleoperation that can identify and complete the user's intended 3D motions

Automated Gait Generation for Walking, Soft Robotic Quadrupeds

Jake Ketchum, T. Murphey

GenerationRobotic Intelligence

🎯 What it does: Designed a sample-efficient, simulation-free, and autonomous method for generating soft robot gaits, demonstrated on a four-legged soft robot with 16 HSA actuators.

Automated Key Action Detection for Closed Reduction of Pelvic Fractures by Expert Surgeons in Robot-Assisted Surgery

Mingzhang Pan, Guibin Bian

ClassificationRobotic IntelligenceConvolutional Neural NetworkRecurrent Neural NetworkTime SeriesBiomedical Data

🎯 What it does: Proposed a multi-task deep learning network architecture combining CNN-BiLSTM and tri-modal fusion for detecting key surgical actions during closed reduction of hip fractures.

Automatic Spatial Radar Camera Calibration via Geometric Constraints with Doppler-Optical Flow Fusion

Jintian Ge, Chen Lv

Autonomous DrivingOptimizationOptical FlowImagePoint Cloud

🎯 What it does: Proposes an automatic spatial calibration method for RGBD cameras and millimeter-wave radar, adopting a two-stage process: first roughly estimating extrinsic parameters using geometric constraints from object contours, then refining them through velocity differences obtained from the camera and radar.

Automotive Radar Missing Dimension Reconstruction from Motion

Chun-Yu Hou, Wen-Chieh Lin

Pose EstimationAutonomous DrivingSimultaneous Localization and MappingPoint Cloud

🎯 What it does: Proposes a method to estimate the pitch angle of stationary targets using the relative speed and radial velocity of automotive radar.

Autonomous Exploration and Mapping for Mobile Robots via Cumulative Curriculum Reinforcement Learning

Z. Li, Ning Li

Robotic IntelligenceReinforcement LearningSimultaneous Localization and Mapping

🎯 What it does: Proposes a Cumulative Curriculum Reinforcement Learning (CCRL) framework that combines deep reinforcement learning with curriculum learning for autonomous exploration and mapping in mobile robots.

Autonomous Exploration Using Ground Robots with Safety Guarantees

David Smith Sundarsingh, Pushpak Jagtap

Safty and PrivacyRobotic Intelligence

🎯 What it does: Proposes an autonomous exploration framework for 2D exploration problems, which takes into account the robot's mathematical model, actual constraints on state and input, as well as hardware limitations, thereby providing safety guarantees for both the robot and the environment.

Autonomous Marker-Less Rapid Aerial Grasping

Erik Bauer, Robert K. Katzschmann

Depth EstimationRobotic IntelligenceConvolutional Neural NetworkPoint Cloud

🎯 What it does: Proposed a vision-based drone rapid grasping system that uses Mask R-CNN segmentation, depth cameras to generate dense point clouds, and geometric grasping planning without relying on markers or known appearances

Autonomous Multi-Robot Servicing for Spacecraft Operation Extension

Longsen Gao, R. Fierro

Robotic IntelligencePhysics RelatedStochastic Differential EquationOrdinary Differential Equation

🎯 What it does: Developed an adaptive robotic system for performing two in-orbit maintenance tasks for spacecraft: customer satellite operations and removal of stuck components (e.g., solar panels), and validated its effectiveness through simulations.

Autonomous Power Line Inspection with Drones via Perception-Aware MPC

Jiaxu Xing, D. Scaramuzza

Object DetectionDomain AdaptationOptimizationRobotic IntelligenceImage

🎯 What it does: Proposed a perception-driven MPC controller for autonomous power line detection and tracking by UAVs, and developed a lightweight detector trained solely on synthetic data with zero-shot transfer to real-world scenarios.

Autonomous Robotic Drilling System for Mice Cranial Window Creation: An Evaluation with an Egg Model

En-dong Zhao, K. Harada

RecognitionRobotic IntelligenceConvolutional Neural NetworkImage

🎯 What it does: Proposed an autonomous robot drilling method for creating skull windows in rodents, including trajectory planning and image-based task completion recognition, validated in eggshell experiments

Autonomous Swarm Robot Coordination via Mean-Field Control Embedding Multi-Agent Reinforcement Learning

Huaze Tang, Xiao-Ping Zhang

Robotic IntelligenceReinforcement Learning

🎯 What it does: Designed a multi-agent reinforcement learning framework called MF-MARL based on mean-field control (MFC) to guide the collaborative behavior of drone swarms.

Autonomous Ultrasound Scanning Towards Standard Plane Using Interval Interaction Probabilistic Movement Primitives

Yi Hu, Mahdi Tavakoli

Robotic IntelligenceConvolutional Neural NetworkBiomedical DataUltrasound

🎯 What it does: Proposed an autonomous ultrasound scanning framework that uses interval probabilistic motion primitives (interval iProMP) to learn scanning navigation strategies from demonstrations, combined with a U-Net to identify target ultrasound images, using confidence maps to evaluate image quality, validated for breast serous cyst scanning.

AV-PedAware: Self-Supervised Audio-Visual Fusion for Dynamic Pedestrian Awareness

Yizhuo Yang, Lihua Xie

Object DetectionAutonomous DrivingRobotic IntelligenceImageMultimodalityAudio

🎯 What it does: Proposed and implemented an audio-visual fusion system called AV-PedAware based on self-supervised learning to enhance pedestrian perception for robots in dynamic environments.

Bag All You Need: Learning a Generalizable Bagging Strategy for Heterogeneous Objects

Arpit Bahety, Shuran Song

Robotic IntelligenceBenchmark

🎯 What it does: Propose a practical robot system that incorporates two learning strategies—rearrangement and lifting—for placing various rigid and deformable objects into deformable bags;

Bagging by Learning to Singulate Layers Using Interactive Perception

L. Chen, Ken Goldberg

Robotic IntelligenceImage

🎯 What it does: Propose SLIP to achieve the separate separation of material layers through interactive perception, and develop the SLIP-Bagging algorithm to complete self-service packaging tasks.

Baking in the Feature: Accelerating Volumetric Segmentation by Rendering Feature Maps

Kenneth Blomqvist, R. Siegwart

SegmentationRepresentation LearningNeural Radiance FieldImage

🎯 What it does: Features extracted using pre-trained models are embedded into NeRF for volumetric segmentation of sparsely annotated color images.

Bang-Bang Boosting of RRTs

Alexander J. LaValle, S. LaValle

OptimizationRobotic Intelligence

🎯 What it does: Proposed a complete and accurate Bang-Bang control method to improve the performance of kinodynamic RRT, and implemented three major applications within RRT;

Beacon-Based Distributed Structure Formation in Multi-Agent Systems

Tamzidul Mina, B. Min

Agentic AI

🎯 What it does: This paper proposes a 3D structural representation method and a distributed formation strategy, enabling fixed agents to guide free-moving agents to land at preset positions, thereby forming a structure.

Bi-Component Silicone 3D Printing with Dynamic Mix Ratio Modification for Soft Robotic Actuators

Brice Parilusyan, Marcos Serrano

OptimizationRobotic Intelligence

🎯 What it does: Propose an additive manufacturing process that utilizes a single two-component silicone rubber with dynamically adjustable mixing ratios to achieve multiple mechanical stiffnesses, and use this process to fabricate multi-channel soft pneumatic actuators, studying the effect of printing direction on bending actuation.

Bi-Level Image-Guided Ergodic Exploration with Applications to Planetary Rovers

Elena Wittemyer, Ian Abraham

OptimizationRobotic IntelligenceConvolutional Neural NetworkImagePhysics Related

🎯 What it does: Proposes an image-based dual-layer entropy quantization exploration method, combining a learned image classifier to update the information map, and optimizing the robot's motion and visual acquisition through coarse and fine solvers to achieve rock formation localization and geological survey by the Mars rover.

Bi-Manual Robot Shoe Lacing

Haining Luo, Y. Demiris

OptimizationRobotic Intelligence

🎯 What it does: Propose a system utilizing a dual-armed robotic system to achieve autonomous shoelace tying, mathematically define the shoelace tying task, search for optimal tying patterns, design action primitives, generate action sequence plans based on patterns, and plan trajectories through active perception of eyelet and shoelace tip positions for execution.

Bidirectional Search Strategy for Incremental Search-based Path Planning

Chenming Li, M. Q. Meng

Robotic Intelligence

🎯 What it does: Proposed the bidirectional incremental search path planning method BLPA*, as well as the fractional bidirectional D* Lite (fBD* Lite(dp)) constrained by the robot's perception range, for collision avoidance path planning in dynamic environments.

Bio-Inspired 3D Flocking Algorithm with Minimal Information Transfer for Drones Swarms

Matthieu Verdoucq, G. Hattenberger

OptimizationRobotic Intelligence

🎯 What it does: A bio-inspired 3D UAV swarm formation algorithm was studied, and its impact on swarm behavior was evaluated in a validated simulation environment.

Bird-View 3D Reconstruction for Crops with Repeated Textures

Guoyu Lu

Depth EstimationImageAgriculture Related

🎯 What it does: Using drone-captured multi-view crop data, propose an unsupervised structure from motion (SfM) framework for large-scale crop 3D reconstruction.

Bistable Tensegrity Robot with Jumping Repeatability Based on Rigid Plate-Shaped Compressors

Kento Shimura, Takuya Umedachi

Robotic IntelligencePhysics Related

🎯 What it does: Developed a bistable tension structure robot capable of performing repetitive jumps using a single motor.

BlinkFlow: A Dataset to Push the Limits of Event-Based Optical Flow Estimation

Yijin Li, Guofeng Zhang

Data SynthesisTransformerOptical FlowBenchmark

🎯 What it does: Propose the BlinkSim simulator to generate large-scale event optical flow data, construct the BlinkFlow dataset, and propose the E-FlowFormer architecture.

Body Posture Controller for Actively Articulated Tracked Vehicles Moving Over Rough and Unknown Terrains

F. Rocha, R. R. Costa

Robotic Intelligence

🎯 What it does: A body posture controller is proposed that utilizes a hinge-based control platform to adjust the platform's posture and clear contact planes when the actively assembled tracked vehicle traverses rough, unknown terrain.

BodySLAM++: Fast and Tightly-Coupled Visual-Inertial Camera and Human Motion Tracking

Dorian Henning, Stefan Leutenegger

Pose EstimationSimultaneous Localization and MappingImage

🎯 What it does: Propose the BodySLAM++ framework, which utilizes visual-inertial data to simultaneously estimate human 6D pose and camera states.

Boosting Feedback Efficiency of Interactive Reinforcement Learning by Adaptive Learning from Scores

Shukai Liu, Liang Zhang

Robotic IntelligenceReinforcement Learning from Human FeedbackReinforcement LearningScore-based Model

🎯 What it does: Improve the feedback efficiency of interactive reinforcement learning by using human-provided scores instead of pairwise preferences, and propose an adaptive learning scheme to handle unstable scores.

Boosting Lidar 3D Object Detection with Point Cloud Semantic Segmentation

Xuchong Zhang, Hongbin Sun

Object DetectionAutonomous DrivingPoint Cloud

🎯 What it does: Propose a multi-task framework that utilizes Cartesian pillars and a multi-scale semantic segmentation head to enhance the performance of 3D object detection using only LiDAR point clouds.

Bootstrapping Adaptive Human-Machine Interfaces with Offline Reinforcement Learning

Jensen Gao, Sergey Levine

Reinforcement Learning

🎯 What it does: Designed and verified an adaptive human-computer interface based on offline reinforcement learning, which can map noisy high-dimensional command signals to actions, and achieve interface learning and improvement through offline pre-training and online fine-tuning.

BRNES: Enabling Security and Privacy-Aware Experience Sharing in Multiagent Robotic and Autonomous Systems

Md Tamjid Hossain, Anton Netchaev

Safty and PrivacyRobotic IntelligenceReinforcement Learning

🎯 What it does: Proposed and implemented the BRNES framework, which reduces the impact of Byzantine attacks by dynamically selecting neighbor regions for each learning agent and adopting weighted experience aggregation techniques; meanwhile, local differential privacy noise is added during the experience sharing process to enhance privacy security.

BSH-Det3D: Improving 3D Object Detection with BEV Shape Heatmap

You Shen, D. Kerr

Object DetectionAutonomous DrivingPoint CloudBenchmark

🎯 What it does: Proposed and implemented the BSH-Det3D model, leveraging BEV shape heatmap to enhance spatial features, and designed the Pillar-based Shape Completion (PSC) module and Attention-based Densification Fusion (ADF) module.

Bubble Explorer: Fast UAV Exploration in Large-Scale and Cluttered 3D-Environments Using Occlusion-Free Spheres

Benxu Tang, Fu Zhang

Optimization

🎯 What it does: Propose a rapid exploration method for unmanned aerial vehicles (UAVs) in large-scale and complex three-dimensional environments

Buoyancy Enabled Non-Inertial Dynamic Walking

Mark Yim, Marc Z. Miskin

Robotic IntelligencePhysics Related

🎯 What it does: Proposes a low Reynolds number pendulum walking mechanism and validates its feasibility through physical demonstrations at two different scales

C2: Co-design of Robots via Concurrent-Network Coupling Online and Offline Reinforcement Learning

Ci Chen, Rong Xiong

Robotic IntelligenceReinforcement Learning

🎯 What it does: Propose a concurrent network framework that combines online and offline reinforcement learning for the co-design of robot morphology and controllers.

Calibration of a Tibia-Based Phase Variable for Control of Robotic Transtibial Prostheses

Ryan R. Posh, Patrick M. Wensing

Robotic IntelligenceTime SeriesBiomedical Data

🎯 What it does: Compared four phase variable calibration methods based on tibial movement to achieve continuous control of a robotic knee prosthesis.

Calibration-Free BEV Representation for Infrastructure Perception

Siqi Fan, Jingjing Liu

Autonomous Driving

🎯 What it does: Proposed a Calibration-free BEV Representation (CBR) network that achieves BEV 3D detection in infrastructure scenarios without relying on camera calibration parameters or additional depth supervision;

CAMETA: Conflict-Aware Multi-Agent Estimated Time of Arrival Prediction for Mobile Robots

Jonas le Fevre Sejersen, Erdal Kayacan

Robotic IntelligenceGraph Neural Network

🎯 What it does: Proposes a conflict-aware multi-agent arrival time prediction framework named CAMETA, which consists of three layers: path planning, ETA prediction, and path selection;

Can Quadruped Guide Robots be Used as Guide Dogs?

Luyao Wang, Jiangtao Gong

Robotic Intelligence

🎯 What it does: Conducted two experiments (indoor controlled scenarios and outdoor real-world scenarios) to evaluate the effectiveness of quadrupedal robots versus wheeled robots in guiding visually impaired and low-vision users.

Canfly: A Can-Sized Autonomous Mini Coaxial Helicopter

Neng Pan, Fei Gao

OptimizationRobotic Intelligence

🎯 What it does: Studied and optimized different rotor configurations, designed a mini dual-rotor drone that is 62% smaller, and proposed its hardware design and control strategies.

CAR-DESPOT: Causally-Informed Online POMDP Planning for Robots in Confounded Environments

Ricardo Cannizzaro, Lars Kunze

Robotic IntelligenceReinforcement Learning

🎯 What it does: Proposed a causally informed extension of AR-DESPOT and designed a method for offline learning of causal model parameters, which was subsequently evaluated on toy problems involving unobserved confounding variables.

CAT-RRT: Motion Planning that Admits Contact One Link at a Time

N. Nechyporenko, Alessandro Roncone

OptimizationRobotic Intelligence

🎯 What it does: Proposed a CAT-RRT planner based on per-link cost heuristics, enabling the robot to complete tasks by traversing high-cost contact regions when necessary.

CDT-Dijkstra: Fast Planning of Globally Optimal Paths for All Points in 2D Continuous Space

Jinyuan Liu, U. Sychou

Optimization

🎯 What it does: Proposed a CDT-Dijkstra algorithm that can rapidly plan globally optimal paths from one point to all points in a 2D continuous space, divided into two stages: SetInit and GetGoal;

CEFHRI: A Communication Efficient Federated Learning Framework for Recognizing Industrial Human-Robot Interaction

Umar Khalid, Chen Chen

RecognitionFederated LearningVideoBenchmark

🎯 What it does: Proposes a communication-efficient federated learning framework called CEFHRI for industrial human-robot interaction, addressing issues of data heterogeneity and communication costs.

Characteristics of Permanent Magnet Coupling Based Wireless Manipulation via Simulation

Tao Zhang, Zheng Li

Robotic IntelligencePhysics Related

🎯 What it does: Evaluated the wireless manipulation characteristics based on permanent magnet coupling, including fixed distance, oscillating torque, and displacement force, and quantified key parameters;

Characterizing the Onset and Offset of Motor Imagery During Passive Arm Movements Induced by an Upper-Body Exoskeleton

Kanishka Mitra, J. D. R. Millán

RecognitionTime SeriesBiomedical Data

🎯 What it does: Studied the initiation and termination characteristics of motor imagery (MI) under passive movement of upper limb exoskeletons, and constructed a decoder using EEG signals to detect transitions from rest to MI initiation and from sustained MI to MI termination;

Chat with the Environment: Interactive Multimodal Perception Using Large Language Models

Xufeng Zhao, Stefan Wermter

Robotic IntelligenceLarge Language ModelVision-Language-Action ModelMultimodality

🎯 What it does: Constructed a robot interaction scenario, proposing an interactive perception framework based on large language models (LLMs). The LLM guides cognitive actions and performs reasoning on multimodal information such as vision, sound, touch, and proprioception to complete task planning.

CineTransfer: Controlling a Robot to Imitate Cinematographic Style from a Single Example

Pablo Pueyo, Mac Schwager

GenerationRobotic IntelligenceVideo

🎯 What it does: Utilizing a single example video, extract shooting style features and achieve video recording that matches this style through a robot.

CLF-CBF Constraints for Real-Time Avoidance of Multiple Obstacles in Bipedal Locomotion and Navigation

Jinze Liu, Jiunn-Kai Huang

OptimizationRobotic IntelligencePoint Cloud

🎯 What it does: Proposed a real-time obstacle avoidance planning system for the Cassie series bipedal robot, utilizing a single differentiable control barrier function (CBF) to achieve obstacle avoidance in environments with multiple non-overlapping obstacles;

CLiFF-LHMP: Using Spatial Dynamics Patterns for Long- Term Human Motion Prediction

Yufei Zhu, Martin Magnusson

Pose EstimationExplainability and InterpretabilitySequential

🎯 What it does: Propose a long-term human motion prediction method called CLiFF-LHMP based on dynamic maps (MoDs), which uses CLiFF-map to bias constant velocity predictions for generating multi-modal trajectory forecasts.

Closed Loop Control of Tendon Driven Continuum Robots Using IMUs

Manu Srivastava, Ian D. Walker

Robotic Intelligence

🎯 What it does: Proposed and implemented a method utilizing IMU quaternion feedback control for continuous soft robotic segments

Closed-Loop Feedback Control of Human Step Width During Walking by Mediolaterally Acting Robotic Hip Exoskeleton

Abbas Alili, H. Huang

Robotic Intelligence

🎯 What it does: Demonstrated a robotic hip exoskeleton utilizing medial-lateral interactions to achieve real-time gait adjustment through admittance control

Co-Speech Gesture Synthesis using Discrete Gesture Token Learning

Shuhong Lu, Andrew W. Feng

GenerationData SynthesisTransformerAuto EncoderMultimodality

🎯 What it does: Propose a two-stage model that synthesizes realistic co-speech gestures through discrete latent codes