检索结果-内蒙古大学图书馆

Performance Improvement of Adaptive N-Square PS-QAM Based on LUT for Flexible Coherent PON

Journal of Lightwave Technology 2025年第12期43卷 5569-5578页

作者： Lin, Gengming Zhu, Yixiao Yan, Zijun Cheng, Ziyu Zhuge, Qunbi Hu, Weisheng Shanghai Jiao Tong University State Key Lab of Advanced Optical Communication Systems and Networks School of Electronic Information and Electrical Engineering Department of Electronic Engineering Shanghai200240 China

Flexible coherent passive optical network (PON) is recognized as an advanced capacity optimization strategy for beyond 100G PON due to its flexibility and adaptability, which are pivotal for the evolution of high-capacity optical access networks. In this study, we propose an adaptive N-square QAM scheme compatible with probabilistic shaping (PS), enabling a more continuous variation of shaping entropy compared to standard 2M -QAM shaping scheme and improved frequency offset estimation (FOE) accuracy in blind 4th-power Fast-Fourier-transform (FFT) algorithm. In experimental demonstrations, dynamic-range and data-rate product (DRDRP) of uniform N2-QAM surpasses that of uniform 2M -QAM by 86 dB·Gbps. The DRDRP of PS N2-QAM also outperforms that of PS 2N -QAM by around 41 dB·Gbps. The development of PON optimization scheme for multi-carrier transmission is achieved through the generation of a look-up table (LUT) based on channel signal-to-noise ratio (SNR) conditions. The feasibility of the LUT is demonstrated under different super-Gaussian filter parameter settings at the back-to-back (B2B) scenario. After 20-km standard single-mode fiber (SSMF) transmission, the normalized generalized mutual information (NGMI) is greater than 0.985 across four subcarriers with varying received optical power (ROP) from −16 to −11 dBm. © 1983-2012 IEEE.

关键词： Single mode fibers

来源：评论

学校读者我要写书评

暂无评论

Learnable digital signal processing:a new benchmark of linearity compensation for optical fiber communications

引用

Light(Science & Applications) 2024年第9期13卷 1931-1943页

作者： Zekun Niu Hang Yang Lyu Li Minghui Shi Guozhi Xu Weisheng Hu Lilin Yi State Key Lab of Advanced Optical Communication Systems and Networks School of Electronic Information and Electrical EngineeringShanghai Jiao Tong UniversityShanghai 200240PR China

The surge in interest regarding the next generation of optical fiber transmission has stimulated the development of digital signal processing(DSP)schemes that are highly cost-effective with both high performance and low *** benchmarks for nonlinear compensation methods,however,traditional DSP designed with block-by-block modules for linear compensations,could exhibit residual linear effects after compensation,limiting the nonlinear compensation *** we propose a high-efficient design thought for DSP based on the learnable perspectivity,called learnable DSP(LDSP).LDSP reuses the traditional DSP modules,regarding the whole DSP as a deep learning framework and optimizing the DSP parameters adaptively based on backpropagation algorithm from a global *** method not only establishes new standards in linear DSP performance but also serves as a critical benchmark for nonlinear DSP *** comparison to traditional DSP with hyperparameter optimization,a notable enhancement of approximately 1.21 dB in the Q factor for 400 Gb/s signal after 1600 km fiber transmission is experimentally demonstrated by combining LDSP and perturbation-based nonlinear compensation *** from the learnable model,LDSP can learn the best configuration adaptively with low complexity,reducing dependence on initial *** proposed approach implements a symbol-rate DSP with a small bit error rate(BER)cost in exchange for a 48%complexity reduction compared to the conventional 2 samples/symbol *** believe that LDSP represents a new and highly efficient paradigm for DSP design,which is poised to attract considerable attention across various domains of optical communications.

关键词： fiber signal optical

来源：评论

学校读者我要写书评

暂无评论

Investigation of system-driven talking approaches using depth and color image information for digital signage with speech user interface 13

Investigation of system-driven talking approaches using dept...

引用

13th IEEE Global Conference on Consumer electronic, GCCE 2024

作者： Takahashi, Kuuta Kawamoto, Shinichi Gunma College Advanced Production Systems Engineering Course Maebashi Japan Gunma College Department of Information and Computer Engineering Maebashi Japan

ISBN: (纸本)9798350355079

In recent decades, digital signages have been widely used. Specifically, digital signages that have a purpose of either providing information or advertising, it is important to direct people's attention toward that information. Conventionally, some approaches to attract viewer's attention include using animation and audible information such as playing background music or a flashing a catchphrase. However, language information has not been implemented in these approaches. In this study, we propose a method to call out a specific person to the front of a digital signage by using color images and depth information. By calling out a specific person by using this method, it is expected for this person to have an awareness that they are a target intended by a digital signage. In this study, we conducted experiments that use utterances based on a person's position, distance from a digital signage, clothing colors, or a pair of two features. Based on the study results, we confirm that a pair of features, that is, clothing color and depth, is the most effective to call out a specific person to the front of a digital signage © 2024 IEEE.

关键词： User interfaces

来源：评论

学校读者我要写书评

暂无评论

Implementation of 400 Gbps quantum noise stream cipher encryption for 1520 km fiber transmission using end-to-end deep learning

引用

Optics Letters 2025年第12期50卷 3808-3811页

作者： Xie, Yunhao Huang, Xianran Xu, Guozhi Xiao, Junzhe Shi, Mengyue Hu, Weisheng Yi, Lilin State Key Laboratory of Advanced Optical Communication Systems and Networks School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai200240 China

In the era of large models and big data, the security of optical fiber communication backbone networks has garnered significant attention. Quantum noise stream cipher (QNSC) stands as a crucial method for safeguarding the physical layer security of optical fiber communications, yet the current schemes lag behind the rate capabilities of existing 400G optical fiber backbone networks. In this paper, we introduce deep learning into QNSC and propose an end-to-end quantum noise stream cipher (E2E-QNSC) scheme, which encrypts 16 quadrature amplitude modulation (QAM) into E2E-65536QAM/QNSC. Our experiments successfully demonstrate secure optical communication with a single-channel rate of 400 Gbps, a total capacity of 8.4 Tbps, and a transmission distance of 1520 km. Even in the most extreme scenarios, the detection failure probability (DFP) of the scheme remains at an excellent level greater than 0.9999, proving the security of the approach. The experimental results presented herein represent the highest rate–distance product record of QNSC secure transmission systems, to the best of our knowledge. © 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

关键词： Quadrature amplitude modulation

来源：评论

学校读者我要写书评

暂无评论

Evolution of Mobile Networks from 4G to 5G: Architectural Transformations New Services and Enhancement of Key Performance Indicators 24

Evolution of Mobile Networks from 4G to 5G: Architectural Tr...

引用

7th International Conference on Networking, Intelligent systems and Security, NISS 2024

作者： Ait Laghchim, Fatima Ezzahra Habibi, Mohamed Mazri, Tomader Electronic Systems Information Processing Mechanical and Energy Systems Ibn Tofail University Morocco Advanced Systems Engineering National School of Applied Sciences Morocco

ISBN: (纸本)9798400709296

The specificity of the mobile network requires providing services to their subscribers with good quality service at all times, 24 hours a day. To achieve this, key performance indicators are defined by radio access standards to verify network performance. In this article, we present the architectures of 4G and 5G networks and their services. We explore the requirements of 5G network services and conduct a study on their key performance. © 2024 ACM.

关键词： 5G mobile communication systems

来源：评论

学校读者我要写书评

暂无评论

Fabric Defect Detection System using YOLO 14

Fabric Defect Detection System using YOLO

引用

14th IIAI International Congress on advanced Applied Informatics, IIAI-AAI 2023

作者： Kawaguchi, Masashi Morimitsu, Yuta Watanabe, Ryusei Ikuta, Chihiro Ishii, Naohiro Umeno, Masayoshi National Institute of Technology Suzuka College Department of Electrical and Electronic Engineering Suzuka Japan Advanced Institute of Industrial Technology Information Systems Architecture Course Tokyo Japan Chubu University Department of Electronic Engineering Kasugai Japan

ISBN: (纸本)9798350324228

In textile processing companies, after purchasing fabric, dyeing, and other treatments, such as surface finishing are carried out before shipping. Defect detection technology is one of the most important issues in this field. Because if we ship a defective fabric without noticing the defect, it will cause damage to the delivery destination. Previously, defect detection of fabric has been performed using methods such as CNN. However, conventional methods such as general multilayer neural networks require two separate processing stages: object detection and identification. Therefore, there is a problem that sufficient real-time performance cannot be guaranteed in the inspection process on an actual manufacturing line. Therefore, in this study, we construct a fabric defect detection system using YOLO which is capable of real-time object detection. In the future, we aim to realize the automation of textile defect detection systems by the construction of an image recognition system using machine learning. It is expected that this will result from reduced detection errors by the human eye, improve detection accuracy, and reduce labor costs. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Solving PnP Problem Using Improved Sparrow Search Algorithm

Solving PnP Problem Using Improved Sparrow Search Algorithm

引用

2024 International Symposium on Intelligent Signal Processing and Communication systems, ISPACS 2024

作者： Riku, Kurihara Matsushima, Kousuke Advanced Engineering Course National Institute of Technology Kurume College Kurume Japan Department of Control and Information Systems Engineering National Institute of Technology Kurume College Kurume Japan

ISBN: (纸本)9798350389210

The Perspective-n-Point (PnP) problem is a fundamental challenge in engineering that plays a crucial role in fields such as computer vision and augmented reality. This problem aims to estimate the position and orientation (pose) of a camera from known 3D points and their corresponding 2D projections on the image plane. With recent technological advancements, application areas such as autonomous vehicles, drones, and robotics are demanding increasingly faster and more accurate camera pose estimation. To address this challenge, we have introduced an improved Sparrow Search Algorithm (SSA) with the capability to dynamically restrict the search area. While the conventional SSA demonstrated excellent optimization capabilities, it faced efficiency issues when searching in high-dimensional spaces, such as those encountered in PnP problems. Our enhancement enables the algorithm to leverage information obtained during the search process to gradually narrow down the search area, thereby more efficiently locating the optimal solution. © 2024 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Mobile Generalized Category Discovery for Road Damage Detection

Mobile Generalized Category Discovery for Road Damage Detect...

引用

2024 International Symposium on Intelligent Signal Processing and Communication systems, ISPACS 2024

作者： Rintaro, Akashi Matsushima, Kousuke Advanced Engineering Course National Institute of Technology Kurume College Kurume Japan Department of Control and Information Systems Engineering National Institute of Technology Kurume College Kurume Japan

ISBN: (纸本)9798350389210

In recent years, there has been a considerable amount of research conducted on the topic of road damage detection using deep learning techniques, with the aim of supporting the safe driving of mobility vehicles. To date, convolutional neural networks and supervised learning with large datasets have been employed for road damage detection, demonstrating high classification accuracy. However, all images in the dataset must be manually labeled. Moreover, images cannot be classified into classes that are not included in the training dataset. The real world presents a vast array of damage types, and the unlabeled input images encompass both known and unknown classes. This renders the system unsuitable for use in providing safe driving support for mobility. Considering this background, Generalized Category Discovery (GCD) has been put forth as a potential solution, aiming to achieve accurate classification in a comprehensive setting where both known and unknown classes are present in unlabeled data. The purpose of this study is to propose a novel model for GCD with the aim of providing safe driving assistance for mobility. The model is based on the pre-trained MobileViT. To validate its performance, this paper examines its performance on the general CIFAR-10 and CIFAR-100 datasets and on a road damage-specific dataset, and discusses the challenges encountered. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Multi-Lens Visual SLAM with Integrated Moving Object Removal using Unsupervised Models 1

Multi-Lens Visual SLAM with Integrated Moving Object Removal...

引用

1st International Conference on Robotics, engineering, Science, and Technology, RESTCON 2024

作者： Hara, Ryoichirou Matsushima, Kousuke Kurume College Advanced Engineering Course National Institute of Technology Fukuoka Japan Kurume College National Institute of Technology Department of Control and Information Systems Engineering Fukuoka Japan

ISBN: (纸本)9798350348842

In advanced Driver-Assistance systems (ADAS), SLAM (Simultaneous Localization and Mapping) technology is required to accurately estimate the position and orientation of onboard cameras. Compared to LiDAR SLAM, Visual SLAM has the advantage of collecting rich environmental information at a lower cost. However, the accuracy of the estimated pose and map construction is greatly reduced by moving objects in the camera scene. To solve this problem, a multi-camera SLAM has been proposed, which works robustly in dynamic environments by using multiple cameras to acquire more feature points from omnidirectional environmental information. In this paper, we propose a robust SLAM system based on a multi-camera SLAM system combined with object detection and semantic segmentation networks. We then evaluate the robustness of the proposed method by comparing it to SLAM systems that do not remove moving objects. Experimental results showed that the Absolute Trajectory Error and the Relative Position Error performed better than conventional SLAM by masking moving objects. It also showed that more static objects were detected and constructed in the map. © 2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Uncertainty-aware Monocular 3D Object Detection for Safe Driving Assistance

Uncertainty-aware Monocular 3D Object Detection for Safe Dri...

引用

2024 International Symposium on Intelligent Signal Processing and Communication systems, ISPACS 2024

作者： Ida, Masahiko Matsushima, Kousuke Advanced Engineering Course National Institute of Technology Kurume College Fukuoka Japan Department of Control and Information Systems Engineering National Institute of Technology Kurume College Fukuoka Japan

ISBN: (纸本)9798350389210

Safety is of utmost importance when utilizing mobile vehicles. The introduction of safe driver assistance and automated driving systems will not only solve this problem, but also reduce road accidents and make driving smoother. To achieve these goals, 3D object detection is an essential technology and various methods have been proposed. Among these, monocular 3D object detection is an important technique and it is very difficult to identify the dimensions, direction and position of an object from a monocular image. Previous methods have predicted the dimensions directly from monocular images, resulting in poor accuracy of the dimensions in the depth direction. However, this paper proposes an indirect method for estimating the dimensions in the depth direction by predicting the volume of 3D bounding box. Detection accuracy for distant and partially visible objects is improved by using a loss function that uses uncertainty to account for the variability in predictions that occurs during learning. The proposed method has received an excellent evaluation on KITTI dataset. © 2024 IEEE.

关键词： 3D modeling

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：