检索结果-内蒙古大学图书馆

Generative Adversarial Networks for Solving Hand-Eye Calibration without Data Correspondence

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Hong, Ilkwon Ha, Junhyoung The RoboticsLAB Hyundai Motor Group Gyeonggi16082 Korea Republic of The Artificial Intelligence and Robotics Institute Korea Institute of Science and Technology Seoul02792 Korea Republic of

In this study, we rediscovered the framework of generative adversarial networks (GANs) as a solver for calibration problems without data correspondence. When data correspondence is not present or loosely established, the calibration problem becomes a parameter estimation problem that aligns the two data distributions. This procedure is conceptually identical to the underlying principle of GAN training in which networks are trained to match the generative distribution to the real data distribution. As a primary application, this idea is applied to the hand-eye calibration problem, demonstrating the proposed method’s applicability and benefits in complicated calibration problems. Copyright © 2024, The Authors. All rights reserved.

关键词： Generative adversarial networks

Real-Time Onboard Object Detection for Augmented Reality: Enhancing Head-Mounted Display with YOLOv8

学校读者我要写书评

暂无评论

Real-Time Onboard Object Detection for Augmented Reality: En...

IEEE International Conference on Edge Computing (EDGE)

作者： Mikołaj Łysakowski Kamil Żywanowski Adam Banaszczyk Michał R. Nowicki Piotr Skrzypczyński Sławomir K. Tadeja Centre for Artificial Intelligence and Cybersecurity Poznań University of Technology Institute of Robotics and Machine Intelligence Poznań University of Technology Department of Engineering Institute for Manufacturing University of Cambridge

This paper introduces a software architecture for real-time object detection using machine learning (ML) in an augmented reality (AR) environment. Our approach uses the recent state-of-the-art YOLOv8 network that runs onboard on the Microsoft HoloLens 2 head-mounted display (HMD). The primary motivation behind this research is to enable the application of advanced ML models for enhanced perception and situational awareness with a wearable, hands-free AR platform. We show the image processing pipeline for the YOLOv8 model and the techniques used to make it real-time on the resource-limited edge computing platform of the headset. The experimental results demonstrate that our solution achieves real-time processing without needing offloading tasks to the cloud or any other external servers while retaining satisfactory accuracy regarding the usual mAP metric and measured qualitative performance.

关键词：

A Novel Framework for Autonomous Landing of the Quadrotor on the Moving Platform by Onboard Vision Sensors 3

学校读者我要写书评

暂无评论

A Novel Framework for Autonomous Landing of the Quadrotor on...

3rd International Conference on Industrial artificial intelligence, IAI 2021

作者： Cao, Qianqian Yu, Hai Fang, Yongchun Liang, Xiao Tianjin Key Laboratory of Intelligent Robotics Nankai University Institute of Robotics and Automatic Information System College of Artificial Intelligence Tianjin China

ISBN: (纸本)9781665435178

This paper addresses the autonomous landing of a quadrotor equipped with sensors and an on-board computer on a moving platform. An autonomous landing framework is proposed, including modules of target detection, state estimation, trajectory planning and tracking control. Specifically, the target detection module utilizes an on-board camera to recognize targets with Apriltags, owing to the high efficiency and accuracy. The state estimation module combines the detected target information and the measurements from the on-board sensors to estimate the absolute motion state of the quadrotor and the platform at the current moment, but also predicts the platform state at the target moment. A dynamically feasible and non-conservative trajectory is generated by the trajectory planning module based on the state estimation. According to the results of the trajectory planning and the state estimation modules, the tracking control module issues control commands to make the quadrotor track the trajectory efficiently. Simulation tests are conducted to validate the feasibility and effectiveness of the proposed framework. © 2021 IEEE.

关键词： State estimation

Point Set Registration Based on Adaptive Maximum Correntropy

学校读者我要写书评

暂无评论

Point Set Registration Based on Adaptive Maximum Correntropy

System Simulation Technology and its Application (CCSSTA), IEEE China Conference on

作者： Yinfeng Shao Wenting Cui Liang Xu School of Electrical Engineering and Automation Tianjin University of Technology Tianjin China Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an China

ISBN: (数字)9798350366600

ISBN: (纸本)9798350366617

Aiming at the problem that the iterative closest point algorithm has a lot of noise in data and poor robustness to outliers, this paper proposes a point set registration algorithm based on the adaptive maximum correntropy criterion. The aim is to enhance the robustness of traditional registration methods, improve registration accuracy, and reduce registration time. Firstly, the maximum correntropy criterion is introduced to effectively identify and eliminate noise and outliers in the data. Secondly, a parameter-adaptive iterative approach is proposed to iteratively remove noise and outliers in the data, thus improving the accuracy and robustness of point set registration. Experimental results on datasets containing a large amount of noise and outliers demonstrate that the proposed algorithm effectively enhances the accuracy and robustness of point set registration.

关键词： Accuracy Adaptive systems Iterative closest point algorithm Noise Birds Robustness Systems simulation Noise measurement Iterative methods Clocks

A Diving Glove with Inertial Sensors for Underwater Gesture Recognition 16th

学校读者我要写书评

暂无评论

A Diving Glove with Inertial Sensors for Underwater Gestur...

16th International Conference on Intelligent robotics and Applications, ICIRA 2023

作者： Tang, Qi Mai, Jingeng Wang, Tiantong Wang, Qining Department of Advanced Manufacturing and Robotics College of Engineering Peking University Beijing100871 China Beijing Engineering Research Center of Intelligent Rehabilitation Engineering Peking University Beijing100871 China Medical Robotics Laboratory University of Health and Rehabilitation Sciences Qingdao266071 China Institute for Artificial Intelligence Peking University Beijing100871 China

ISBN: (纸本)9789819964857

Underwater gesture recognition has emerged as a popular research area for achieving efficient and secure underwater human-human interaction and human-robot collaboration. Previous research primarily relied on visual methods, which face challenges related to low visibility and motion blur in underwater environments. In this paper, we introduce a diving glove embedded with inertial sensors for underwater gesture recognition. The Nearest Centroid (NC), Random Forest (RF) and Support Vector Machine (SVM) classifiers are used in our diving glove. To demonstrate the underwater gesture recognition performance, we conducted a two-stage underwater experiment with ten underwater gestures. For the same user, The average recognition accuracies of the NC, RF and SVM classifiers are 97.5 % ± 5.4 %, 98.6 % ± 0.6 % and 99.2 % ± 0.3 %, respectively. For new users, the average recognition accuracies of the NC, RF and SVM classifiers are 86.4 % ± 2.18 %, 88.6 % ± 7.4 % and 96.5 % ± 0.31 %, respectively. The study suggests that the diving glove with inertial sensors is a feasible solution for underwater gesture recognition. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023.

关键词： Gesture recognition

Learning-Augmented Scheduling for Solar-Powered Electric Vehicle Charging

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Li, Tongxin Sun, Chenxi The Chinese University of Hong Kong Shenzhen China the Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China

We tackle the challenge of learning to charge Electric Vehicles (EVs) with Out-of-Distribution (OOD) data. Traditional scheduling algorithms typically fail to balance near-optimal average performance with worst-case guarantees, particularly with OOD data. Model Predictive Control (MPC) is often too conservative and data-independent, whereas Reinforcement Learning (RL) tends to be overly aggressive and fully trusts the data, hindering their ability to consistently achieve the best-of-both-worlds. To bridge this gap, we introduce a novel OOD-aware scheduling algorithm, denoted OOD-CHARGING. This algorithm employs a dynamic "awareness radius", which updates in real-time based on the Temporal Difference (TD)-error that reflects the severity of OOD. The OOD-CHARGING algorithm allows for a more effective balance between consistency and robustness in EV charging schedules, thereby significantly enhancing adaptability and efficiency in real-world charging environments. Our results demonstrate that this approach improves the scheduling reward reliably under real OOD scenarios with remarkable shifts of EV charging behaviors caused by COVID-19 in the Caltech ACN-Data. © 2023, CC BY.

关键词： Scheduling algorithms

A Hybrid Algorithm Based on Ant Colony and Genetic Algorithm for AUV Path Planning

学校读者我要写书评

暂无评论

A Hybrid Algorithm Based on Ant Colony and Genetic Algorithm...

Fully Actuated System Theory and Applications (CFASTA), Conference on

作者： Jihong Hu Meng Joo Er Tianhe Liu Siqi Wang Institute of Artificial Intelligence and Marine Robotics College of Marine Electrical Engineering Dalian Maritime University Dalian China

In this paper, a hybrid algorithm based on ant colony optimization (ACO) and genetic algorithm(GA) is proposed for path planning of an autonomous underwater vehicle. The novel algorithm, which combines ACO and GA has the following features: ❨1❩ It plans path through ACO; ❨2❩ GA is used to find the optimal path; ❨3❩ A new evaluation function is proposed and is adjusted to select a relatively smooth path; ❨4❩ The number of search points is reduced to reduce computation load. Simulation results show that the proposed algorithm can find shorter paths in the same map compared with ACO.

关键词：

Grasp What You Want: Embodied Dexterous Grasping System Driven by Your Voice

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Junliang Ye, Kai Kang, Haolan Liang, Mingxuan Wu, Yuhang Liu, Zhenhua Zhuang, Huiping Huang, Rui Chen, Yongquan Shenzhen Institute of Artificial Intelligence and Robotics for Society China The Chinese University of Hong Kong Shenzhen China South China University of Technology China

In recent years, as robotics has advanced, human-robot collaboration has gained increasing importance. However, current robots struggle to fully and accurately interpret human intentions from voice commands alone. Traditional gripper and suction systems often fail to interact naturally with humans, lack advanced manipulation capabilities, and are not adaptable to diverse tasks, especially in unstructured environments. This paper introduces the Embodied Dexterous Grasping System (EDGS), designed to tackle object grasping in cluttered environments for human-robot interaction. We propose a novel approach to semantic-object alignment using a Vision-Language Model (VLM) that fuses voice commands and visual information, significantly enhancing the alignment of multi-dimensional attributes of target objects in complex scenarios. Inspired by human hand-object interactions, we develop a robust, precise, and efficient grasping strategy, incorporating principles like the thumb-object axis, multi-finger wrapping, and fingertip interaction with an object’s contact mechanics. We also design experiments to assess Referring Expression Representation Enrichment (RERE) in referring expression segmentation, demonstrating that our system accurately detects and matches referring expressions. Extensive experiments confirm that EDGS can effectively handle complex grasping tasks, achieving stability and high success rates, highlighting its potential for further development in the field of Embodied AI. © 2024, CC BY-NC-SA.

关键词： Human robot interaction

MIM4D: Masked Modeling with Multi-View Video for Autonomous Driving Representation Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zou, Jialv Liao, Bencheng Zhang, Qian Liu, Wenyu Wang, Xinggang School of EIC Huazhong University of Science & Technology China Institute of Artificial Intelligence Huazhong University of Science & Technology China Horizon Robotics China

Learning robust and scalable visual representations from massive multi-view video data remains a challenge in computer vision and autonomous driving. Existing pre-training methods either rely on expensive supervised learning with 3D annotations, limiting the scalability, or focus on single-frame or monocular inputs, neglecting the temporal information. We propose MIM4D, a novel pre-training paradigm based on dual masked image modeling (MIM). MIM4D leverages both spatial and temporal relations by training on masked multi-view video inputs. It constructs pseudo-3D features using continuous scene flow and projects them onto 2D plane for supervision. To address the lack of dense 3D supervision, MIM4D reconstruct pixels by employing 3D volumetric differentiable rendering to learn geometric representations. We demonstrate that MIM4D achieves state-of-the-art performance on the nuScenes dataset for visual representation learning in autonomous driving. It significantly improves existing methods on multiple downstream tasks, including BEV segmentation (8.7% IoU), 3D object detection (3.5% mAP), and HD map construction (1.4% mAP). Our work offers a new choice for learning representation at scale in autonomous driving. Code and models are released at https://***/hustvl/MIM4D. Copyright © 2024, The Authors. All rights reserved.

关键词： Autonomous vehicles