检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Liang, Wenyu Baldivieso, Pablo R. Drummond, Ross Shin, Donghwan Dept. of Automatic Control and Systems Engineering University of Sheffield Sheffield United Kingdom Dept. of Computer Science University of Sheffield Sheffield United Kingdom

Typical autonomous driving systems are a combination of machine learning algorithms (often involving neural networks) and classical feedback controllers. Whilst significant progress has been made in recent years on the neural network side of these systems, only limited progress has been made on the feedback controller side. Often, the feedback control gains are simply passed from paper to paper with little re-tuning taking place, even though the changes to the neural networks can alter the vehicle’s closed loop dynamics. The aim of this paper is to highlight the limitations of this approach;it is shown that retuning the feedback controller can be a simple way to improve autonomous driving performance. To demonstrate this, the PID gains of the longitudinal controller in the TCP autonomous vehicle algorithm are tuned. This causes the driving score in CARLA to increase from 73.21 to 77.38, with the results averaged over 16 driving scenarios. Moreover, it was observed that the performance benefits were most apparent during challenging driving scenarios, such as during rain or night time, as the tuned controller led to a more assertive driving style. These results demonstrate the value of developing both the neural network and feedback control policies of autonomous driving systems simultaneously, as this can be a simple and methodical way to improve autonomous driving system performance and robustness. © 2024, CC BY.

关键词： Feedback control

来源：评论

学校读者我要写书评

暂无评论

Error Bounds for Decoding Piecewise Constant Nanopore Signals in DNA Storage

Error Bounds for Decoding Piecewise Constant Nanopore Signal...

引用

IEEE International Conference on Communications (ICC)

作者： Adrian Vidal V. B. Wijekoon Emanuele Viterbo Dept. Electrical & Computer Systems Engineering Monash University Clayton VIC Australia

Nanopore sequencing enables reading strings of A,C,G,T nucleotides in DNA strands by pulling them into nanopores with the help of motor proteins. Due to the discrete stepping of motor proteins, the signals produced by a DNA sequence tend to be piecewise-constant expansions of some underlying real-valued sequence. In this paper, we assume that every k-nucleotide sequence corresponds to a real-valued codeword of length $k$ , and model the nanopore channel as a noisy duplication channel that stretches every sample of a codeword using a geometric distribution, and then adds Gaussian noise. We show that for this channel, a simpler variant of the dynamic time warping (DTW) algorithm performs maximum likelihood decoding. Next, we devise an $O(k^{2})$ - algorithm for bounding the pairwise error probability between two codewords of length $k$ . Finally, we use Scrappie to design codebooks with a storage efficiency of 1 bit per nucleotide and demonstrate using error simulations the accuracy of the calculated error bounds.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose Estimation and Motion Inbetweening

ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose E...

引用

IEEE Workshop on Applications of computer Vision (WACV)

作者： Hojun Jang Young Min Kim Dept. of Electrical and Computer Engineering Seoul National University Interdisciplinary Program in Artificial Intelligence and INMC Seoul National University

ISBN: (数字)9798331510831

ISBN: (纸本)9798331510848

We present Reusable Motion prior (ReMP), an effective motion prior that can accurately track the temporal evolution of motion in various downstream tasks. Inspired by the success of foundation models, we argue that a robust spatio-temporal motion prior can encapsulate underlying 3D dynamics applicable to various sensor modalities. We learn the rich motion prior from a sequence of complete parametric models of posed human body shape. Our prior can easily estimate poses in missing frames or noisy measurements despite significant occlusion by employing a temporal attention mechanism. More interestingly, our prior can guide the system with incomplete and challenging input measurements to quickly extract critical information to estimate the sequence of poses, significantly improving the training efficiency for mesh sequence recovery. ReMP consistently outperforms the baseline method on diverse and practical 3D motion data, including dept. point clouds, LiDAR scans, and IMU sensor data. Project page is available in https://***/ReMP.

关键词： Training Point cloud compression Three-dimensional displays Accuracy Tracking Shape Pose estimation Noise Parametric statistics Noise measurement

来源：评论

学校读者我要写书评

暂无评论

Union Bound for Generalized Duplication Channels with DTW Decoding

Union Bound for Generalized Duplication Channels with DTW De...

引用

IEEE International Symposium on Information Theory

作者： Adrian Vidal V. B. Wijekoon Emanuele Viterbo Dept. Electrical & Computer Systems Engineering Monash University Clayton VIC Australia

In this paper, we calculate a union bound for dynamic time warping (DTW)-based decoding of piecewise constant signals corrupted by additive noise and time stretching due to sample duplications, as observed in raw measurement signals obtained from nanopore sequencers. We consider both finitely- and infinitely-supported duplications with geometric-like characteristic, which include discrete uniform distributions as a special case. First, we provide explicit algorithms that calculate the union bound in O(αk 2 ) time for the infinite-support case and in O(β 2 k 2 ) for the finite-support case, where k is the codeword length, α is the minimum duplication, and β is the maximum duplication. Next, we show that a multi-read union bound exhibits a thresholding effect, where the error probability can be made arbitrarily close to zero by aggregating DTW distances from sufficiently many independent reads. Finally, we validate the calculated bounds relative to simulation results.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Hydrological and Structural Monitoring for Flood Risk Mitigation in Cosenza Urban Area (Southern Italy)

Hydrological and Structural Monitoring for Flood Risk Mitiga...

引用

International Conference on Information and Communication Technologies for Disaster Management (ICT-DM)

作者： Daniela Biondi Francesco S. Liguori Antonio Bilotta Antonio Madeo Giovanna Capparelli Dept. of Computer Engineering Modelling Electronics and Systems University of Calabria Rende Italy

The evaluation of flood risk in urban areas characterized by high vulnerability indices is a fundamental task in order to undertake any mitigation action. When flood-induced phenomena affect structures (buildings, bridges, river banks, etc.) particularly in areas adjacent to rivers, as for the city of Cosenza, the real time detection of any change in the structural response is crucial for a proper safety level assessment. In the present work, a framework for remote monitoring of hydrological events and related structural behaviour is proposed. Data pertaining to the hydro-meteorological event are collected mainly through a series of ultrasonic hydrometers, different types of rain gauges and webcams, to allow a rough estimate of the expected action on the structures. The structural response is monitored through accelerometers (or other instruments) which allows identifying the evolution of the dynamic behaviour, in order to detect insurgent damage phenomena. The acquired structural measurements are employed to tune both fine structural models or rough simplified models. Moreover, the structural monitoring system allows to estimate damages due to other natural events, such as earthquakes, which well suits current philosophies of multi-risk assessment. This study is carried out within the “Smart integrated monitoring system for safe and resilient communities” action of the NRRP Tech4you project.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CPR: Correlation-based Page Remapping

CPR: Correlation-based Page Remapping

引用

International SoC Design Conference, ISOCC

作者： Hojung Namkoong Jungrae Kim Dept. of Electrical and Computer Engineering Sungkyunkwan University Suwon Korea Dept. of Semiconductor Systems Engineering Sungkyunkwan University Suwon Korea

As the number of multimedia IPs in mobile devices increases, a high memory bandwidth becomes essential. To meet this demand, Application Processors (APs) employ multiple memory channels. However, employing fine-grained channel interleaving leads to frequent bank conflicts in DRAM, resulting in increased energy consumption and reduced battery life. This paper proposes a novel Virtual to Physical address mapping (VA-to-PA) scheme, called Correlation-based Page Remapping (CPR). This technique involves remapping pages based on the access correlations between them to enhance bank parallelism. By increasing the channel interleaving size to 1KB and applying CPR, we can reduce the DRAM energy consumption of activation (ACT) and precharge (PRE) operations by 15% without any decrease in DRAM performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Systematic Literature Review of the Application of Agile Applied to Large-Scale, Safety-Critical, Cyber-Physical systems

A Systematic Literature Review of the Application of Agile A...

引用

International Conference on Data and Software engineering (ICODSE)

作者： Robin J. Yeman Yashwant K. Malaiya Dept. Systems Engineering Colorado State University Fort Collins USA Dept. Computer Science Colorado State University Fort Collins USA

ISBN: (数字)9798331506407

ISBN: (纸本)9798331506414

The application of Agile methodologies to large-scale, safety-critical cyber-physical systems (LS/SC/CPS) has shown significant interest over the last 5 years. Although there has been limited research into each of the individual attributes for de-velopment projects that are large-scale, safety-critical, or cyber-physical systems, there is a scarcity of research when all three attributes exist simultaneously in a development endeavor. This systematic literature review explores the extent of Agile adoption in projects where all three attributes are present. It begins by examining the extent of Agile adoption by domain, identifying the drivers for change, reviewing the challenges encountered, and investigating the augmentations used to address these challenges. The findings indicate that there has been growth in adoption across the medical, aerospace, automotive, and robotics indus-tries, with limited adoption in energy and transportation. Key drivers include speed, adaptability, and cost. Several challenges have been identified, with regulatory/compliance requirements being the most significant. Multiple augmentations to Agile have been applied to overcome the identified challenges, ranging from incorporating safety frameworks and specialized artifacts to leveraging modeling and simulation, as well as reimagining pro-grammatics. This review highlights industry efforts to apply Agile to LS/SC/CPS, while identifying several gaps in the literature. The largest gap is empirical validation in real-world settings. Future research should focus on improving the existing Agile scaling frameworks with guidance and resources to help achieve the benefits of Agile in LS/SC/CPS development while ensuring regulatory and safety compliance.

关键词： Industries Technological innovation Costs Transportation Cyber-physical systems Complexity theory Standards Aerospace industry Systematic literature review Software engineering

来源：评论

学校读者我要写书评

暂无评论

Investigating the Significance of Surface Roughness for Screen-printed Silver ink on Kapton Substrate

Investigating the Significance of Surface Roughness for Scre...

引用

IISE Annual Conference and Expo 2022

作者： Abdelatty, Mohamed Y. Umar, Ashraf Khinda, Gurvinder S. Poliks, Mark D. Dept. of Systems Science and Industrial Engineering State University of New York at Binghamton BinghamtonNY United States Dept. of Electrical and Computer Engineering State University of New York at Binghamton BinghamtonNY United States

ISBN: (纸本)9781713858072

Minor changes between the designed Flexible hybrid electronic (FHE) devices and the actual realized printed devices are very crucial for the development and maturity of this technology. The quality of printed electronics depends on mechanical and chemical factors such as ink properties and the printing technique [1], [2]. Fluctuations in the quality of printed electronic devices, such as thickness, width, waviness, and surface roughness, affect their radio frequency (RF) performance [2]. In this paper, we investigate statistically, using design of experiments (DoE) techniques, the significance of surface roughness of screen-printed silver ink traces on the RF performance. Ninety-six microstrip lines were printed;we measured the average thickness, average width, average roughness, and average waviness, for all the samples. Also, we measured the S-parameters for all the samples;then, we extracted Insertion-loss at four different frequencies. We performed simple statistical analysis and checked independence and normality assumptions for our data. We considered the thickness, width, roughness, and waviness to be our independent variables. We performed regression analysis and best subsets regression four times each time we consider the insertion-loss as the dependent variable. We were able to conclude that surface roughness becomes more significant as the frequency increases for screen-printed FHEs. © 2022 IISE Annual Conference and Expo 2022. All rights reserved.

关键词： Surface roughness

来源：评论

学校读者我要写书评

暂无评论

Utilizing Deep Neural Networks to Improve Intrusion Detection System(IDS)

Utilizing Deep Neural Networks to Improve Intrusion Detectio...

引用

Telecommunications Conference (ITC-Egypt), International

作者： Ahmed Mahmoud Saad Mohammed Awad Meekay Mostafa Samy El Sayed Mansour Electronics and Computer and Systems Engineering Dept. Mansoura University Cairo Egypt Mansour Electronics and Communications Dept. Mansoura University Cairo Egypt

ISBN: (数字)9798350351408

ISBN: (纸本)9798350351415

Internet services are now accessible everywhere in the world due to the rapid advancement of networks and technology in recent decades. The development of information security solutions that can recognize novel threats is now critical due to the surge in piracy and the compromising of many modern systems. An intrusion detection system (IDS) is one of the most important information security technologies. It uses machine learning and deep learning algorithms to locate anomalies in the network. The main goal of this paper is to identify the unknown assault package by utilizing a novel intrusion detection method that offers excellent network performance. In this model, two techniques-binary classification and multiclass classification-that make use of a deep neural network algorithm are used to detect attacks. The suggested's outstanding accuracy

关键词： Deep learning Machine learning algorithms Accuracy Web and internet services Supervised learning Intrusion detection Artificial neural networks

来源：评论

学校读者我要写书评

暂无评论

Multispectral Image Segmentation in Agriculture: Evaluating Deep Learning Models with Train-Test Split and Cross-Validation Strategies

Multispectral Image Segmentation in Agriculture: Evaluating ...

引用

Iberian Robotics Conference (ROBOT)

作者： Wilgo Cardoso Tiago Barros Gil Gonçalves Cristiano Premebida Urbano J. Nunes Department of Electrical and Computer Engineering Institute of Systems and Robotics University of Coimbra Portugal Dept. of Mathematics Institute for Systems Engineering and Computers at Coimbra University of Coimbra Portugal

ISBN: (数字)9798350376364

ISBN: (纸本)9798350376371

In agricultural robotics, the integration of multi-spectral image processing and deep learning (DL) has become the state-of-the-art (SOTA) in crop monitoring, yield estimation, and efficient land management. This work addresses the impact of different DL-segmentation models and evaluation protocols on multispectral imagery datasets collected by a UAV over vineyards. In terms of evaluation protocols, we have considered train-test split, standard k-fold cross-validation, and group k-fold cross-validation. While the first two assume that the training and test data are drawn from the same underlying distribution, the group k-fold cross-validation protocol assumes that each fold represents distinct distributions. Most works either adopt train-test split or k-fold cross-validation under the assumption that both the training and test sets are drawn from the same distribution. However, this assumption is rarely met in real-world applications. Therefore, the objective of this study is to evaluate and compare different evaluation protocols within the context of a real-world agricultural task, highlighting their limitations and weaknesses. Two SOTA DL-based segmentation models, SegNet and DeepLabV3, are employed to perform semantic segmentation on datasets of three Vineyards. The models have been trained and tested considering single-modality representations. In addition to the RGB modality, models trained on NDVI, GNDVI and early fusion are also evaluated. The performance of the models are evaluated using the IoU metric across different dataset configurations. The results indicate that the early fusion representation achieves the highest performance across the various splitting protocols, compared to the single-input representations. The results also show that the train-test and random k-fold splitting approaches report similar results. However, when employing group k-fold the performance drops consistently across both models and the modalities. This indicates that the models

关键词： Deep learning Training Precision agriculture Measurement Protocols Data models Yield estimation Robots Monitoring Standards

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：