检索结果-内蒙古大学图书馆

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Wang, Zhuxin Wang, Juan Xia, Chengyi Tianjin Univ Technol Tianjin Key Lab Control Theory & Complicated Ind Sch Elect & Elect Engn Tianjin 300384 Peoples R China Tiangong Univ Sch Artificial Intelligence Tianjin 300384 Peoples R China

ISBN: (纸本)9798350321050

Knee meniscus tear is a common joint disorder that is usually diagnosed with the help of MRI imaging. However, the diagnosis of a meniscal tear places high technical demands on physicians, and the results of the diagnosis are also inconsistent. In this paper, a meniscal tear detection method based on YOLOv5 target detection network is developed to help physicians make more accurate diagnoses. To begin, a channel and space parallel attention module is designed and integrated into the feature fusion part of the network to improve the network's attention to the tear area. Then, ConvNeXt is used in the backbone network part to improve the C3 module to obtain the ConvC3 module, which strengthens the ability of the backbone network to extract the features of the meniscal tear lesion area. After labelling and creating the dataset, the improved YOLOv5 network is trained to obtain the target model. The experimental results show that compared with the original model, the improved YOLOv5 model has an increased mAP from 82.5% to 84.8%, a slight decrease in GFLOPs as well, and presents an overall considerable improvement. This model plays an important role in the diagnosis of meniscal tears.

关键词： Knee meniscus tear MRI imaging YOLOv5 Target detection network

来源：评论

学校读者我要写书评

暂无评论

FedVPS: Federated learning for Privacy and Security of Internet of Vehicles on Non-IID data 12

FedVPS: Federated Learning for Privacy and Security of Inter...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Kuang Hangdong Mi Bo Huang Darong Deng Zhaoyang Chongqing Jiaotong Univ Coll Informat Sci & Engn Chongqing 400074 Peoples R China Anhui Univ Sch Artificial Intelligence Hefei 230039 Anhui Peoples R China

ISBN: (纸本)9798350321050

The 6th generation of mobile networks (6G) offers many advantages, including fast transmission rates, low delays, and ultra-dense networks. These benefits can solve the communication problems in the Internet of Vehicles (IoV). However, automobiles as IoV terminals generate Non-Independent Identically Distributed (Non-IID) data while driving, making it necessary to introduce distributed machine learning into the IoV as a network that integrates intelligent computing and vehicle networking. To address the Non-IID of local data and the heterogeneity of local models, we propose a heterogeneous Federated learning scheme in this paper. Based on the distributed architecture (Terminal - Edge device - Cloud) in the IoV, we designed a privacy protection scheme using Secure Multi-Party Computation (SMPC). This ensures that the terminals participating in Federated learning get accurate calculation results without revealing useful information, thus preserving the privacy of local datasets. The privacy and security of the IoV based on Federated learning (FedVPS) not only protect the privacy of the terminals but also improve communication efficiency, enabling accurate and efficient distributed machine learning. The aggregation method of Federated learning is a prototype-based scheme that utilizes the effective information stored in local datasets. In the BIT-Vehicle dataset, FedVPS is not only more robust but also has excellent prediction accuracy. Compared to FedAvg, FedVPS has advantages in communication efficiency and model prediction accuracy.

关键词： Internet of Vehicles Non-IID SMPC FedVPS

来源：评论

学校读者我要写书评

暂无评论

data-driven System Identification of Quadrotors Subject to Motor Delays

Data-Driven System Identification of Quadrotors Subject to M...

引用

2024 International conference on Intelligent Robots and systems

作者： Eschmann, Jonas Albani, Dario Loianno, Giuseppe NYU Tandon Sch Engn Brooklyn NY 11201 USA Technol Innovat Inst Autonomous Robot Res Ctr Abu Dhabi U Arab Emirates

ISBN: (纸本)9798350377712;9798350377705

Recently non-linear control methods like Model Predictive control (MPC) and Reinforcement learning (RL) have attracted increased interest in the quadrotor control community. In contrast to classic control methods like cascaded PID controllers, MPC and RL heavily rely on an accurate model of the system dynamics. The process of quadrotor system identification is notoriously tedious and is often pursued with additional equipment like a thrust stand. Furthermore, low-level details like motor delays which are crucial for accurate end-to-end control are often neglected. In this work, we introduce a data-driven method to identify a quadrotor's inertia parameters, thrust curves, torque coefficients, and first-order motor delay purely based on proprioceptive data. The estimation of the motor delay is particularly challenging as usually, the RPMs can not be measured. We derive a Maximum A Posteriori (MAP)-based method to estimate the latent time constant. Our approach only requires about a minute of flying data that can be collected without any additional equipment and usually consists of three simple maneuvers. Experimental results demonstrate the ability of our method to accurately recover the parameters of multiple quadrotors. It also facilitates the deployment of RL-based, end-to-end quadrotor control of a large quadrotor under harsh, outdoor conditions.

关键词： Predictive control systems

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Supervised learning-Based Model-Free Adaptive Fault-Tolerant Spacecraft Attitude control With Deferred Asymmetric Constraints

引用

ieee TRANSACTIONS ON AEROSPACE AND ELECTRONIC systems 2023年第6期59卷 8884-8900页

作者： Sun, Xiaoyun Shen, Qiang Wu, Shufan Shanghai Jiao Tong Univ Sch Aeronaut & Astronaut Shanghai 200240 Peoples R China

Aiming at solving the model-free fault-tolerant spacecraft attitude control problem, a data-driven adaptive control scheme is proposed to the spacecraft in the presence of actuator faults and performance constraints. First, the discrete-time attitude dynamic model is transformed into an affine nonlinear system based on local dynamic linearization. Then, a fuzzy logic-based lazily adapted constant kinky inference rule is introduced to predict the arbitrarily continuous nonlinear actuator faults and model uncertainties by supervised learning with insufficient prior knowledge. To satisfy time-varying deferred asymmetric constraints of the attitude tracking error, a virtual control law is proposed in the attitude control loop using the back-stepping approach, which is derived from a deferred switching transformation and barrier Lyapunov function. The stability of the data-driven adaptive fault-tolerant attitude control for the nonlinear discrete-time spacecraft attitude dynamics is analyzed rigorously with the aid of contraction mapping principle and discrete-time Lyapunov theory. Compared with existing methods, the proposed one considers nominal nonglobal Lipschitz nonlinear system with arbitrarily continuous unmodeled uncertainties and time-varying actuator faults, and achieves smaller prediction error bound than general kinky inference scheme and better closed-loop performance by estimating and compensating unknown dynamics. Finally, numerical simulation verifies the effectiveness of the proposed control scheme.

关键词： Space vehicles Attitude control Actuators Adaptation models Uncertainty Aerodynamics Fault tolerant systems Adaptive control barrier Lyapunov function contraction mapping principle data-driven control fault-tolerant control (FTC) kinky inference spacecraft attitude control supervised learning

来源：评论

学校读者我要写书评

暂无评论

Research on target grab of leg-arm cooperative robot based on vision 12

Research on target grab of leg-arm cooperative robot based o...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Xia, Haichuang Zhang, Xiaoping North China Univ Technol Sch Elect & Control Engn Beijing 100144 Peoples R China Nanjing Univ Aeronaut & Astronaut Coll Automat Engn Nanjing 211106 Peoples R China Univ Cambridge Dept Engn Cambridge CB2 1PZ England

ISBN: (纸本)9798350321050

Leg-arm cooperative robot can not only expand the motion space of traditional manipulator, but also make up the operation ability of traditional mobile robot. This paper studies the target grab of leg-arm cooperative robot based on vision. First of all, independently develop the leg-arm cooperative robot. Secondly, for the operation of the manipulator, a height-angle uncoupling PID algorithm is innovatively designed in the cylindrical coordinate system. The target position information obtained by the YOLOv3 algorithm is used as feedback to realize the accurate control of the manipulator, which can avoid the complex process of trajectory planning. Furthermore, combined with three-element trajectory planning method and motion control scheme under triangular gait, a target grab control strategy for the leg-arm cooperative robot is designed, including the height allocation decision maker and the target tracking controller. Finally, the proposed algorithm and schemes are verified with the entity robot.

关键词： leg-arm cooperative control hexapod robot manipulator arm YOLOv3 target grab

来源：评论

学校读者我要写书评

暂无评论

learning Optimal Antenna Tilt control Policies: A Contextual Linear Bandits Approach

引用

ieee TRANSACTIONS ON MOBILE COMPUTING 2024年第12期23卷 12666-12679页

作者： Vannella, Filippo Proutiere, Alexandre Jedra, Yassir Jeong, Jaeseong Ericsson Res S-16483 Stockholm Sweden KTH Royal Inst Technol Div Decis & Control Syst S-11428 Stockholm Sweden MIT Lab Informat & Decis Syst Cambridge MA 02139 USA

controlling antenna tilts in cellular networks is critical to achieve a good trade-off between network coverage and capacity. We devise algorithms learning optimal tilt control policies from existing data (passive learning setting) or from data actively generated by the algorithms (active learning setting). We formalize the design of such algorithms as a Best Policy Identification problem in Contextual Linear Bandits (CLB). In CLB, an action represents an antenna tilt update;the context captures current network conditions;the reward corresponds to an improvement of performance, mixing coverage and capacity. The objective is to identify an approximately optimal policy (a function mapping the context to an action with maximal reward). For both active and passive learning, we derive information-theoretical lower bounds on the number of samples required by any algorithm returning an approximately optimal policy with a given level of certainty, and devise algorithms achieving these fundamental limits. We apply our algorithms to the Remote Electrical Tilt optimization problem in cellular networks, and show that they can produce optimal tilt update policy using much fewer data samples than naive or existing rule-based learning algorithms. This paper is an extension of work presented at ieee International conference on Computer Communications (INFOCOM) 2022 (Vannella et al. 2022).

关键词： Antennas Optimization Approximation algorithms Interference Mobile computing control systems Context modeling Antenna tilt optimization coverage capacity optimization best policy identification contextual linear bandits

来源：评论

学校读者我要写书评

暂无评论

Iterative learning control for Two-Sensors System with User's Preference 13

Iterative Learning Control for Two-Sensors System with User'...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Zhang, Zhenfa Shen, Dong School of Mathematics Renmin University of China Beijing100872 China Engineering Research Center of Finance Computation and Digital Engineering Ministry of Education China

ISBN: (纸本)9798350361674

This study proposes an iterative learning control (ILC) scheme for a two-sensor system with user's preference. ILC updates the system input using error information from previous iterations to sequentially enhance tracking performance. Typ-ically, it is challenging to minimize the tracking performances of both sensors simultaneously. Therefore, we aim to track the Pareto optimal inputs that correspond to different trade-offs in tracking performances. In certain applications, there may be a requirement for the tracking performances of both sensors to satisfy a specific proportionality relationship, i.e., user's preference. In this study, we introduce an algorithm that addresses the tracking problem in a two-sensor system while considering user's preference. This algorithm not only tracks a Pareto optimal input but also ensures that the tracked performances meet user's preference. © 2024 ieee.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

controllable Blind AC FDIA via Physics-Informed Extrapolative AVAE

引用

SENSORS 2025年第3期25卷 943-943页

作者： Zhao, Siliang Luo, Wuman Shu, Qin Xu, Fangwei Sichuan Univ Coll Elect Engn Chengdu 610000 Peoples R China Macao Polytech Univ Sch Appl Sci Macau Peoples R China

False data injection attacks (FDIAs) targeting AC state estimation pose significant challenges, especially when only power measurements are available, and voltage measurements are absent. Current machine learning-based approaches struggle to effectively control state estimation errors and are confined to the data distribution of training sets. To address these limitations, we propose the physics-informed extrapolative adversarial variational autoencoder (PI-ExAVAE) for generating controllable and stealthy false data injections. By incorporating physically consistent priors derived from the AC power flow equations, which enforce both the physical laws of power systems and the stealth requirements to evade bad data detection mechanisms, the model learns to generate attack vectors that are physically plausible and stealthy while inducing significant and controllable deviations in state estimation. Experimental results on ieee-14 and ieee-118 systems show that the model achieves a 90% success rate in bypassing detection tests for most attack configurations and outperforms methods like SAGAN by generating smoother, more realistic deviations. Furthermore, the use of physical priors enables the model to extrapolate beyond the training data distribution, effectively targeting unseen operational scenarios. This highlights the importance of integrating physics knowledge into data-driven approaches to enhance adaptability and robustness against evolving detection mechanisms.

关键词： controllable false data injection attack AC state estimation physics informed data driven extrapolative adversarial variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Finite-Time Annular Domain Stability Analysis of Switched Positive Linear systems 13

Finite-Time Annular Domain Stability Analysis of Switched Po...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Yan, Zhiguo Tang, Zhenyu Zhu, Baolong Jinan250353 China

ISBN: (纸本)9798350361674

This article investigates the issue of finite-time annular domain stability (FTADS) of switched positive linear systems (SPLSs), particularly focusing on the incorporation of the concept of average dwell time (ADT). By using the multiple linear copositive Lyapunov function (MLCLF), a sufficient condition for FTADS of SPLSs is derived. The derived condition is formulated as a linear programming problem, facilitating numerical verification. The proposed result is substantiated through two numerical examples. © 2024 ieee.

关键词： Lyapunov functions

来源：评论

学校读者我要写书评

暂无评论

Adaptive Prescribed Performance control for Nonlinear Multiagent systems with Disturbance 13

Adaptive Prescribed Performance Control for Nonlinear Multia...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Chen, He Cao, Liang Pan, Yingnan College of Control Science and Engineering Bohai University Jinzhou121013 China College of Mathematical Sciences Bohai University Jinzhou121013 China

ISBN: (纸本)9798350361674

This paper is dedicated to studying the prescribed performance control (PPC) problem for nonlinear multiagent systems with saturation and unknown disturbance. A filter that integrates second-order sliding mode techniques is designed to reduce the computation burden. According to the approximation characteristic of neural networks (NNs), the disturbance observer is designed to compensate the unknown disturbance. With PPC technology, the tracking error converges to a certain range. Different from the existing results, this paper examines the saturation issue that exists in the first step of the virtual controller. NNs are utilized to approximate the unknown saturation function. Furthermore, by exploiting Lyapunov stability theory, all closed-loop system signals are ensured to be semiglobally uniformly ultimately bounded. Finally, the rationality of the presented strategy is confirmed by the numerical simulation. © 2024 ieee.

关键词： Adaptive control systems

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：