检索结果-内蒙古大学图书馆

5th International Academic Exchange Conference on Science and technology Innovation, IAECST 2023

作者： Fan, Hong Wang, Haixiang Che, Kai Wu, Ziyu Aeronautics Computing Technology Research Institute Pre-research Department Xian China Aeronautics Computing Technology Research Institute Laboratory 18th Xian China

ISBN: (纸本)9798350357738

Hardware accelerators for deep learning algorithms are the cornerstone of intelligence in the aviation field. The adoption of Field Programmable Gate Array (FPGA) as embedded coprocessors for algorithm acceleration has gained wide acceptance. Decision-making games are critical tasks, where their accuracy and real-time performance directly influence the outcome of aerial engagements. Traditional FPGA development using hardware description languages is associated with long development cycles, high complexity, slow iteration, and challenges in swiftly responding to model algorithm updates. In contrast, FPGA design based on High Level Synthesis (HLS) provides an appropriate technology pathway to circumvent these drawbacks. A specific deep learning algorithm related to aircraft decision games is implemented, optimized, and deployed on FPGA via HLS. The FPGA is connected to the host machine via PCIe as a coprocessor, achieving holistic coupled acceleration. Comparative testing against running the algorithm on an ftd2000 eight-core CPU and an Intel i9-9900k processor demonstrates that using FPGA as a coprocessor significantly reduces algorithm execution time, resulting in a noticeable acceleration effect. © 2023 IEEE.

关键词： High level synthesis

来源：评论

学校读者我要写书评

暂无评论

A Review of GPU Virtualization technology Based on API Redirection 5

A Review of GPU Virtualization Technology Based on API Redir...

引用

5th International Academic Exchange Conference on Science and technology Innovation, IAECST 2023

作者： Zhao, Yuhong Gao, Ze Wu, Ziyu Xu, Ling Aeronautics Computing Technology Research Institute Pre-research Department Xian China Aeronautics Computing Technology Research Institute Laboratory 18th Xian China

ISBN: (纸本)9798350357738

With the continuous development of cloud computing, artificial intelligence and other technologies, the importance of GPU virtualization technology has been continuously strengthened in deep learning and intelligent computing. Compared with traditional device pass-through technology, quasi-virtualization technology and full virtualization technology, virtualization technology based on API redirection not only achieves resource segmentation, but also does not need to obtain detailed equipment documents and driver source code, which has great advantages. GPU virtualization technology based on API redirection is mainly based on user-mode API redirection and kernel-mode API redirection, which are the two mainstream development directions. At present, there are already implementation schemes, which have certain differences and connections in terms of time-sharing, space-sharing, computing isolation and other aspects. At the same time, the future development of GPU virtualization technology should pay more attention to the aspects of preemption and security. © 2023 IEEE.

关键词： Virtualization

来源：评论

学校读者我要写书评

暂无评论

Att2CPC: Attention-Guided Lossy Attribute Compression of Point Clouds

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2025年第3期6卷 639-650页

作者： Liu, Kai You, Kang Gao, Pan Paul, Manoranjan Nanjing University of Aeronautics and Astronautics College of Computer Science and Technology Nanjing211106 China Charles Sturt University School of Computing and Mathematics BathurstNSW2678 Australia

With the great progress of three-dimensional (3-D) sensing and acquisition technology, the volume of point cloud data has grown dramatically, which urges the development of efficient point cloud compression methods. In this article, we focus on the task of learned lossy point cloud attribute compression (PCAC). We propose an efficient attention-based method for lossy compression of point cloud attributes leveraging on an autoencoder architecture. Specifically, at the encoding side, we conduct multiple downsampling to best exploit the local attribute patterns, in which effective external cross attention (ECA) is devised to hierarchically aggregate features by intergrating attributes and geometry contexts. At the decoding side, the attributes of the point cloud are progressively reconstructed based on the multiscale representation and the zero-padding upsampling tactic. To the best of our knowledge, this is the first approach to introduce attention mechanism to point-based lossy PCAC task. We verify the compression efficiency of our model on various sequences, including human body frames, sparse objects, and large-scale point cloud scenes. Experiments show that our method achieves an average improvement of 1.15 and 2.13 dB in Bjontegaard delta (BD)-peak signal-to-noise ratio (BD-PSNR) of Y channel and YUV channel, respectively, when comparing with the state-of-the-art point-based method deep-PCAC. Codes of this article are available at https://***/I2-Multimedia-Lab/Att2CPC. © 2024 IEEE.

关键词： HTTP

来源：评论

学校读者我要写书评

暂无评论

Research on Concurrent Destruction of Embedded High-Availability Distributed computing Platform 24

Research on Concurrent Destruction of Embedded High-Availabi...

引用

24th IEEE/ACIS International Conference on Computer and Information Science, ICIS 2024

作者： He, Qinhan Liu, Jinhui He, Lijun Zhou, Guangfei School of Artificial Intelligence Xidian University Xi'an China School of Computer Science and Technology Xidian University Xi'an China Aeronautics Computing Technique Research Institute Xi'an China

ISBN: (纸本)9798350376647

The embedded distributed computing platform runs large-scale tasks with high complexity and high computing capacity requirements, and divides the embedded tasks required by super computing capacity into multiple independent sub-tasks, which can be simultaneously allocated to multiple computing nodes for paralleled execution. This paper researches that the embedded high-availability distributed computing platform ensures the real-time and deterministic requirements of embedded system function execution through the distribution of computing resources, network communication resources and storage resources, and implements the allocation and execution of distributed tasks under static partition configuration based on the distributed computing framework of the Actor model. In the paper, the destruction technology of large-scale storage resources in embedded high-availability distributed computing platform is also studied, the concurrent destruction is proposed, which improves the efficiency of destructive power input, solves the problems of large amount of computing nodes destruction, large destructive current and long destructive duration of large-scale storage resources in embedded system, and effectively solves the leakage of key information. © 2024 IEEE.

关键词： Concurrent destruction Distributed computing High availability

来源：评论

学校读者我要写书评

暂无评论

Research on Complex Electronic System Health Management technology 2

Research on Complex Electronic System Health Management Tech...

引用

2nd International Conference on Electronic Information Engineering and Computer Communication, EIECC 2022

作者： Yan, Tanbao Niu, Wei Wang, Sen Gao, Han Aeronautics Computing Technology Research Institute Research Department Xi'an*** China

ISBN: (纸本)9781510663121

At present, the structure of electronic system is more and more complex, and the function is more and more important. However, the degradation state of electronic products has no obvious external performance, and there are sudden and random faults. The development of fault diagnosis technology of electronic products is relatively slow. In view of the above problems, this paper proposes a complex electronic system health management method, which defines the modules such as state management, execution management, NVRAM manager and watchdog manager to define the functions and cross-linking relationships of each component, which realizes the health management of complex electronic system under the joint supervision of software application and hardware electronic equipment. © 2023 SPIE.

关键词： Oscillators (electronic)

来源：评论

学校读者我要写书评

暂无评论

Model Compression Method Based on Knowledge Distilling and Adversarial Learning 1

引用

14th International Conference on Complex Systems Design and Management, CSD and M 2023

作者： Du, Ming Ma, Zhenhua Liu, Yuan Xi’an Aeronautics Computing Technology Research Institute Xi’an710068 China

ISBN: (数字)9789819965113

ISBN: (纸本)9789819965106

For airborne computing platform, fault diagnosis model is a very important part of Prognostics and Health Management System. Because of the lack of labels, unsupervised learning is commonly used in fault diagnosis model. However, unsupervised learning method requires too much computing power and storage, so it’s hard to be deployed to airborne computing platform with limited computing power and storage. In this paper, we proposed a model compression method named Optimized Adversarial Distilling Model Compression (OADMC) that combines Knowledge Distilling and Adversarial Learning to solve this problem. In our method, we trained an Artificial Neural Network (ANN) model to imitate the unsupervised model, and use Simulated Annealing (SA) method to find the difference between them. OADMC can made required computing power reduced to less than 25% and required storage reduced to less than 5% and the accuracy is higher than 95%, whose performance is better than traditional model compression such as pruning algorithm. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： computing power

来源：评论

学校读者我要写书评

暂无评论

The Theoretical Analysis based on Effective Capacity in Massive MIMO-NOMA Systems 16

The Theoretical Analysis based on Effective Capacity in Mass...

引用

16th International Conference on Communication Software and Networks, ICCSN 2024

作者： Yanmei, Cao Guomei, Zhang Xuping, Qi Jun, Li Jun, Zhao Lin, Yan Xi'An Aeronautics Computing Technology Rasearch Institute Xi'An China Xi'An Jiaotong University Xi'An China

ISBN: (纸本)9798331506322

In this paper, for the delay-sensitive service requirements, an effective capacity model considering the delay QoS exponent is introduced in massive MIMO-NOMA system. We derive the theoretical expression of the effective capacity in the two-user scenario within one cluster for the massive MIMONOMA system, and analyze the impact of the SNR(Signal-toNoise Ratio) and delay QoS(Quality of Service) exponent on the effective capacity. The simulation results show that theoretical analysis results is consistent with the Monte Carlo simulation results within the allowable error range. Secondly, some factors that affect the effective capacity are analyzed, including the influence of the delay QoS exponent and the transmission SNR on the effective capacity. The simulation results verify the correctness of the theoretical analysis, which provide an important theoretical basis for the next research. © 2024 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Resource Allocation Scheme based on Effective Capacity in Massive MIMO-NOMA Systems 16

Resource Allocation Scheme based on Effective Capacity in Ma...

引用

16th International Conference on Communication Software and Networks, ICCSN 2024

作者： Yanmei, Cao Guomei, Zhang Lin, Wang Jun, Zhao Jun, Li Xi'An Aeronautics Computing Technology Rasearch Institute Xi'An China Xi'An Jiaotong University Xi'An China

ISBN: (纸本)9798331506322

With the explosive growth of mobile data and the exponential growth of IoT applications, the communication field has increased its demand for high data rates and low latency in various aspects. In massive MIMO-NOMA systems, research on resource allocation mostly focuses on maximizing system capacity or minimizing transmission power. This cannot fully meet the high-speed and low latency requirements in future communication scenarios. In response to this issue, this paper takes the effective capacity constrained by the QoS index as the resource optimization objective, in order to maximize the system's effective capacity while ensuring the QoS requirements. The paper aims to maximize the effective capacity and designs a power allocation algorithm based on bisection and a user clustering scheme based on DQN. The simulation results show that the proposed algorithm can effectively improve the effective capacity of the system. © 2024 IEEE.

关键词： Constrained optimization

来源：评论

学校读者我要写书评

暂无评论

An Enumeration-Based Deterministic Algorithm for Solving optimization Problem 6

An Enumeration-Based Deterministic Algorithm for Solving opt...

引用

6th IEEE International Conference on Automation, Electronics and Electrical Engineering, AUTEEE 2023

作者： Yang, Changlin Aeronautics Computing Technology Research Institute Process Management Office Xi'an China

ISBN: (纸本)9798350305623

An enumeration-based deterministic algorithm is proposed for solving optimization problem. Discretization of the feasible region for each decision variable is applied to establish search tree so that all the solutions could be enumerated. A new search region is established from the optimization of previous literation. Meanwhile, adjustment of search step with exponential decay function is conducive to reducing the search scope until the convergence of result is realized. The proposed algorithm is applied for three kinds of typical optimization problem, including Non-linear Programing Problem, Job-Shop Scheduling Problem and Traveling Salesman Problem. The results obtained by the new algorithm are compared with results of swarm intelligent optimization algorithm. Results indicate that for the combinational optimization, high convergence rate could be realized by adjusting the feasible region for each decision variable. For non-linear programming problem, the accuracy of results obtained from deterministic algorithm is better than that from self-organizing migrating algorithm, which infers that enumeration-based deterministic algorithm could be used to solve NP hard problem. © 2023 IEEE.

关键词： deterministic algorithm enumeration job-shop scheduling problem nonlinear programming problem optimization traveling salesman problem

来源：评论

学校读者我要写书评

暂无评论

Research on the Application of Multi-Source Radar Information Fusion Based on Digital Twin 7

Research on the Application of Multi-Source Radar Informatio...

引用

7th International Conference on Computer Information Science and Application technology, CISAT 2024

作者： Zhao, Sheng Xie, Yanan Lu, Tegang Bai, Hao Northwest Institute of Mechanical and Electrical Engineering Xianyang China Aeronautics Computing Technology Research Institute Xian China

ISBN: (纸本)9798350375107

This paper proposes a multi-source radar information fusion technology grounded in digital twin technology, considering the future development trajectories of radar information fusion and the current status and advancements in digital twin technology. It systematically examines the specific applications of digital twin technology in multi-source radar information fusion, encompassing system architecture design, data acquisition and processing, real-time simulation and prediction, fault diagnosis, and health management. Additionally, the potential applications of digital twin technology in fault diagnosis and health management within the context of multi-source radar information fusion are explored, proposing data-driven preventive maintenance strategies to further enhance system reliability and maintainability. The paper also identifies challenges such as high-precision modeling, data transmission, and processing efficiency in the application of digital twin technology within multi-source radar information fusion, providing reference directions for future development. © 2024 IEEE.

关键词： Data fusion

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：