检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Jena, Rohit Sethi, Deeksha Chaudhari, Pratik Gee, James C. Computer and Information Science Electrical and Systems Engineering Radiology Penn Image Computing and Science Laboratory United States

Classical optimization and learning-based methods are the two reigning paradigms in deformable image registration. While optimization-based methods boast generalizability across modalities and robust performance, learning-based methods promise peak performance, incorporating weak supervision and amortized optimization. However, the exact conditions for either paradigm to perform well over the other are shrouded and not explicitly outlined in the existing literature. In this paper, we make an explicit correspondence between the mutual information of the distribution of per-pixel intensity and labels, and the performance of classical registration methods. This strong correlation hints to the fact that architectural designs in learning-based methods is unlikely to affect this correlation, and therefore, the performance of learning-based methods. This hypothesis is thoroughly validated with state-of-the-art classical and learning-based methods. However, learning-based methods with weak supervision can perform high-fidelity intensity and label registration, which is not possible with classical methods. Next, we show that this high-fidelity feature learning does not translate to invariance to domain shift, and learning-based methods are sensitive to such changes in the data distribution. Finally, we propose a general recipe to choose the best paradigm for a given registration problem, based on these observations. © 2024, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Forward-looking Sonar Based Autonomous Aquaculture Inspection

Forward-looking Sonar Based Autonomous Aquaculture Inspectio...

引用

OCEANS

作者： Daniel Rosa David Cabecinhas Fausto Ferreira Institute for Systems and Robotics Instituto Superior Técnico Lisbon LABUST Faculty of Electrical Engineering and Computing Zagreb

ISBN: (数字)9798350362077

ISBN: (纸本)9798350362084

In recent years, there has been a rising interest in aquaculture development due to the increasing demand for food driven by population's growth. Aquaculture infrastructures require regular monitoring to assure their preservation and food safety, traditionally conducted by slow and risky manual methods involving professional divers. To solve this issue, an autonomous aquaculture net inspection using an Autonomous Underwater Vehicle (AUV) is proposed. The AUV is equipped with a Forward-Looking Sonar (FLS), which provides acoustic images of the environment. By using image processing techniques to remove noise, the vehicle's distance and orientation relative to the net cage are measured. The net cage's center is estimated, using a circle regression, to compute the relative orientation. Then, velocity and orientation references are generated, so the AUV can cover the net whole perimeter, while staying at a desired distance. Finally, many tests are performed to analyze the vehicle's behavior while encircling the net.

关键词： Autonomous underwater vehicles Current measurement Sonar measurements Measurement uncertainty Sonar Sea measurements Inspection

来源：评论

学校读者我要写书评

暂无评论

An Efficient Wireless Channel Estimation Model for Environment Sensing

arXiv

引用

arXiv 2024年

作者： Zaidi, Zainab R. Alpcan, Tansu Leckie, Christopher Erfani, Sarah Department of Electrical and Electronic Engineering The University of Melbourne Australia School of Computing and Information Systems The University of Melbourne Australia

This paper presents a novel and efficient wireless channel estimation scheme based on a tapped delay line (TDL) model of wireless signal propagation, where a data-driven machine learning approach is used to estimate the path delays and gains. The key motivation for our novel channel estimation model is to gain environment awareness, i.e., detecting changes in path delays and gains related to interesting objects and events in the field. The estimated channel state provides a more detailed measure to sense the field than the single-tap channel state indicator (CSI) in current OFDM systems. Advantages of this approach also include low computation time and training data requirements, making it suitable for environment awareness applications. We evaluate this model’s performance using Matlab’s ray-tracing tool under static and dynamic conditions for increased realism instead of the standard evaluation approaches that rely on classical statistical channel models. Our results show that our TDL-based model can accurately estimate the path delays and associated gains for a broad-range of locations and operating conditions. Root-mean-square estimation error was less than 10−4, or −40dB, for SNR ≥ 60dB in all of our experiments. Our results show that interference of a flying drone on signal multipaths, in a preliminary experiment, can be detected in estimated channel states which, otherwise, remains obscured in conventional CSI. © 2024, CC BY.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

AI-Enhanced Structural Health Monitoring with a Multi-Rotor Aerial Vehicle

AI-Enhanced Structural Health Monitoring with a Multi-Rotor ...

引用

International Conference on Unmanned Aircraft systems (ICUAS)

作者： Filip Zorić Ana Milas Tamara Petrović Zdenko Kovačić Matko Orsag Faculty of Electrical Engineering and Computing LARICS Laboratory for Robotics and Intelligent Control Systems University of Zagreb Zagreb Croatia

ISBN: (数字)9798350357882

ISBN: (纸本)9798350357899

Roads, bridges, tunnels, railways and canals are crucial for the transportation of goods and people in modern society. Modern infrastructure is subject to damage and requires therefore regular inspection and maintenance to remain safe and functional. While this can be done manually, large parts of the mentioned infrastructure are hardly accessible, which is a main motivational factor for the use of multi-rotor aerial vehicles (MAV s) for inspection and maintenance tasks. MAV s, paired with novel deep learning computer vision techniques, such as instance segmentation, are emerging as an obvious choice for automating certain parts of structural health monitoring (SHM). In this manuscript, we provide an overview of available SHM datasets and how they can be utilized for the task of autonomous crack detection. We use available datasets to train and compare two neural network architectures for instance segmentation. The segmented crack instances are then localized in a global coordinate frame to perform autonomous mapping of the potentially dangerous infrastructure defects. Experimental studies demonstrate the effectiveness of the proposed SHM approach, which results in precisely localized infrastructure defects marked on the global map.

关键词： Instance segmentation Irrigation Structural panels Neural networks Transportation Inspection Rail transportation

来源：评论

学校读者我要写书评

暂无评论

Generating Visual Stimuli from EEG Recordings using Transformer-encoder based EEG encoder and GAN

arXiv

引用

arXiv 2024年

作者： Mishra, Rahul Bhavsar, Arnav Multimedia Analytics Networks and Systems Lab. School of Computing & Electrical Engineering IIT Mandi Mandi India

In this study, we tackle a modern research challenge within the field of perceptual brain decoding, which revolves around synthesizing images from EEG signals using an adversarial deep learning framework. The specific objective is to recreate images belonging to various object categories by leveraging EEG recordings obtained while subjects view those images. To achieve this, we employ a Transformer-encoder based EEG encoder to produce EEG encodings, which serve as inputs to the generator component of the GAN network. Alongside the adversarial loss, we also incorporate perceptual loss to enhance the quality of the generated images. © 2024, CC0.

关键词： Gallium nitride

来源：评论

学校读者我要写书评

暂无评论

A Distance-Geometric Method for Recovering Robot Joint Angles From an RGB Image

引用

IFAC-PapersOnLine 2023年第2期56卷 1003-1008页

作者： Ivan Bilić Filip Marić Ivan Marković Ivan Petrović University of Zagreb Faculty of Electrical Engineering and Computing Laboratory for Autonomous Systems and Mobile Robotics University of Toronto Space and Terrestrial Autonomous Robotic Systems Laboratory

Autonomous manipulation systems operating in domains where human intervention is difficult or impossible (e.g., underwater, extraterrestrial or hazardous environments) require a high degree of robustness to sensing and communication failures. Crucially, motion planning and control algorithms require a stream of accurate joint angle data provided by joint encoders, the failure of which may result in an unrecoverable loss of functionality. In this paper, we present a novel method for retrieving the joint angles of a robot manipulator using only a single RGB image of its current configuration, opening up an avenue for recovering system functionality when conventional proprioceptive sensing is unavailable. Our approach, based on a distance-geometric representation of the configuration space, exploits the knowledge of a robot's kinematic model with the goal of training a shallow neural network that performs a 2D-to-3D regression of distances associated with detected structural keypoints. It is shown that the resulting Euclidean distance matrix uniquely corresponds to the observed configuration, where joint angles can be recovered via multidimensional scaling and a simple inverse kinematics procedure. We evaluate the performance of our approach on real RGB images of a Franka Emika Panda manipulator, showing that the proposed method is efficient and exhibits solid generalization ability. Furthermore, we show that our method can be easily combined with a dense refinement technique to obtain superior results.

关键词： Manipulation Mechatronic systems Robotics Joint angle estimation

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning for Joint Resource and Power Allocation in D2D Communications

Reinforcement Learning for Joint Resource and Power Allocati...

引用

IEEE Conference on Decision and Control

作者： Ifrah Saeed Andrew C. Cullen Zainab R. Zaidi Sarah Erfani Tansu Alpcan Department of Electrical and Electronic Engineering The University of Melbourne Australia School of Computing and Information Systems The University of Melbourne Australia

ISBN: (数字)9798350316339

ISBN: (纸本)9798350316346

Device-to-Device (D2D) communication is critical in many public safety scenarios where access to network infrastructure is not guaranteed. In such out-of-coverage situations, random resource allocation is a straightforward way for direct communication by each D2D pair. However, such a resource allocation plan often results in severe interference and low throughput, which can degrade crucial communications performance. In contrast, the problem of centralised joint resource block selection and power control optimisation to maximise the total data rate of all D2D pairs, known as the sum-rate, is non-convex and NP-hard but can effectively reduce interference. To address the inability to solve this non-convex centralised problem setting due to limited information in resource-constrained out-of-coverage scenarios, we propose two distributed reinforcement learning schemes. Our methods allow D2D pairs to autonomously make decisions on their joint resource block and power allocation to minimise their mutual interference and maximise their sum-rate, while maintaining the quality of service constraints. To evaluate these methods in out-of-coverage scenarios, we conduct extensive performance evaluations using the ns-3 network simulator designed for public safety LTE-D2D. We demonstrate that our algorithms, in the absence of any centralised coordination and neighbouring information, autonomously reach time-averaged sum-rates that are within $98.2 \%$ and $98.6 \%$ of the rates achieved by the centralised optimisation solution.

关键词： Wireless communication Training Power control Interference Quality of service Reinforcement learning Public security Device-to-device communication Resource management Optimization

来源：评论

学校读者我要写书评

暂无评论

Dynamic AI-Driven Network Slicing with O-RAN for Continuous Connectivity in Connected Vehicles and Onboard Consumer Electronics

引用

IEEE Transactions on Consumer Electronics 2025年第1期71卷 720-733页

作者： Shah, Syed Danial Ali Bashir, Ali Kashif Al-Otaibi, Yasser D. Dabel, Maryam M. Al Ali, Farman University of Leeds School of Electronic and Electrical Engineering LeedsLS2 9JT United Kingdom Manchester Metropolitan University Department of Computing and Mathematics M15 6BH United Kingdom Chitkara University Institute of Engineering and Technology Chitkara University Centre for Research Impact and Outcome Punjab Rajpura140401 India King Abdulaziz University Faculty of Computing and Information Technology in Rabigh Department of Information Systems Jeddah21589 Saudi Arabia University of Hafr Al Department of Computer Science and Engineering Batin Saudi Arabia Sungkyunkwan University School of Convergence Seoul Korea Republic of

The rise of connected and autonomous vehicles signifies an era of intelligent transportation systems, where robust and continued network connectivity is essential for critical applications and enhanced in-vehicle Consumer Electronics (CE) experiences. Slicing at the network's edge offers tailored and dedicated logical networks for diverse and low-latency vehicular demands, including Advanced Driver Assistance systems (ADAS) and in-car infotainment. However, seamless migration of network slices as vehicles traverse coverage areas of different network operators presents formidable challenges, such as ensuring continuous connectivity and uninterrupted service for both safety-critical systems and consumer-oriented services. In this paper, we introduced dynamic network slicing for continuous connectivity in connected vehicles and onboard CE using the Open Radio Access Network (O-RAN) framework in a highly dynamic and mobile environment. We implemented an xAPP within O-RAN that enables Deep Reinforcement Learning (DRL) agent to learn optimal policies through interaction with the network, guiding intelligent decisions on slice migration, resource allocation, and handover optimization. We conducted simulations and evaluations to demonstrate the effectiveness of the proposed xAPP in maintaining optimal Quality of Service (QoS), ensuring efficient RAN resource utilization, minimizing service interruptions, and prioritizing safety-critical slices, all while supporting seamless operation of CE within vehicles during mobility. © 1975-2011 IEEE.

关键词： Advanced driver assistance systems

来源：评论

学校读者我要写书评

暂无评论

Low Back Muscle Fatigue Monitoring Using Proprietary electrical Impedance Myography System

Low Back Muscle Fatigue Monitoring Using Proprietary Electri...

引用

2023 IEEE Biomedical Circuits and systems Conference, BioCAS 2023

作者： Xu, Pan Culjak, Ivana Roglic, Matija Klaic, Luka Gao, Yueming Vasic, Zeljka Lucev Fuzhou University College of Physics and Information Engineering Fuzhou350108 China Fuzhou University The International Joint Laboratory on Health Intelligent Monitoring Systems Fuzhou350108 China University of Zagreb Faculty of Electrical Engineering and Computing Zagreb10000 Croatia

ISBN: (纸本)9798350300260

With the change of modern lifestyle, the number of people living a sedentary life is increasing, which can lead to sustained tension in the low back muscles, which in turn can cause lumbar muscle strain and even low back pain. Measuring low back muscle fatigue in real time could help prevent and rehabilitate low back pain and improve the health of sedentary individuals. In this paper we aim to estimate muscle fatigue of low back muscles in vivo using electrical impedance myography (EIM). EIM measurements were performed using a proprietary impedance measuring device and the commercial ImpTMSFB7 human body composition analyzer. Impedances measured with both devices showed correlation greater than 0.99 and a linear decreasing trend with duration of fatigue contraction. The impedance of the lumbar muscles measured at the end of the experiment when fatigue occurred (240 s) was 16±2 Ω lower than the impedance measured at the beginning of the experiment (0 s). The measured resistance decreased by approximately 13 % and the reactance increased by approximately 8.5 % during low back muscle contraction, indicating a sufficient sensitivity of the muscle impedance graph in detecting low back muscle fatigue. These results contribute to the research of EIM assessment of low back muscle fatigue and provide valuable guidance for additional research. © 2023 IEEE.

关键词： electrical impedance myography (EIM) low back muscles muscle fatigue rehabilitation monitoring

来源：评论

学校读者我要写书评

暂无评论

Grid Frequency Control Capability of Energy Storage systems: Modeling, New Control Approach, and Real-time Validation 8

Grid Frequency Control Capability of Energy Storage Systems:...

引用

8th IEEE Workshop on the Electronic Grid, eGRID 2023

作者： Oshnoei, Arman Oshnoei, Soroush Jalilpoor, Kamran Soudjani, Sadegh Blaabjerg, Frede Aalborg University Department of Energy Aalborg Denmark Aarhus University Department of Electrical and Computer Engineering Aarhus Denmark Ghent University Department of Electromechanical Systems and Metal Engineering Ghent Belgium Newcastle University School of Computing Newcastle United Kingdom

ISBN: (纸本)9798350327007

Energy storage systems (ESSs) have proved to be efficient in frequency regulation by providing flexible charging/discharging powers. This paper presents a model predictive control (MPC) with feedback correction (FC) to provide the ESS with control signals to be efficiently involved in the frequency regulation in a power system with renewable power generation. The FD is introduced to improve the accuracy of the prediction in the MPC. An approach based on the artificial neural network (ANN) is presented for optimal design of the weighting coefficients appearing in the MPC objective function. The controller performance is compared with an MPC without feedback correction, a fuzzy-PD control, and a scheme with no support from the ESS. A comparison is also made to examine the effect of weighting coefficients tuned by the ANN with those tuned by a fuzzy intelligent method and a sine-cosine algorithm. Real-time validations are provided to demonstrate the proposed method's effectiveness. © 2023 IEEE.

关键词： Model predictive control

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：