检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Sun, Haoran Huang, Bangchao Zhang, Zishang Xu, Ronghan Huang, Guojing Feng, Shihao Huang, Guangyi Yin, Jiayi Qiu, Nuofan Chen, Hua Zhang, Wei Pan, Jia Wan, Fang Song, Chaoyang Department of Mechanical and Energy Engineering Southern University of Science and Technology China Department of Computer Science The University of Hong Kong Hong Kong School of Design Southern University of Science and Technology China ZJU-University of Illinois Urbana-Champaign Institute Zhejiang University China System Design and Intelligent Manufacturing Southern University of Science and Technology China Design & Learning Research Group Southern University of Science and Technology China

This paper studies the design, control, and learning of a novel robotic limb that produces overconstrained locomotion by employing the Bennett linkage for motion generation, capable of parametric reconfiguration between a reptile- and mammal-inspired morphology within a single quadruped. In contrast to the prevailing focus on planar linkages, this research delves into adopting overconstrained linkages as the limb mechanism. The overconstrained linkages have solid theoretical foundations in advanced kinematics but are under-explored in robotic applications. This study showcases the morphological superiority of Overconstrained Robotic Limbs (ORLs) that can transform into planar or spherical limbs, exemplified using the simplest case of a Bennett linkage as an ORL. We apply Model Predictive Control (MPC) to simulate a range of overconstrained locomotion tasks, revealing its superiority in energy efficiency against planar limbs when considering foothold distances and speeds. The results are further verified in overconstrained locomotion policies optimized from Reinforcement Learning (RL). From an evolutionary biology perspective, these findings highlight the mechanism distinctions in limb design between reptiles and mammals and represent the first documented instance of ORLs outperforming planar limb designs in dynamic locomotion. Future studies will focus on deploying the model-based and learning-based overconstrained locomotion skills in the robotic hardware to close the Sim2Real gap for developing evolutionary-inspired, energy-efficient control of novel robotic limbs. © 2023, CC BY-NC-ND.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning-based Token Pruning in Vision Transformers: A Markov Game Approach

arXiv

引用

arXiv 2025年

作者： Lu, Chenglong Liang, Shen Wang, Xuewei Wang, Wei School of Computer Science Fudan University China LIPADE Université Paris Cité Paris France State Key Laboratory of Mechanical Behavior and System Safety of Traffic Engineering Shijiazhuang Tiedao University Shijiazhuang China

Vision Transformers (ViTs) have computational costs scaling quadratically with the number of tokens, calling for effective token pruning policies. Most existing policies are handcrafted, lacking adaptivity to varying inputs. Moreover, they fail to consider the sequential nature of token pruning across multiple layers. In this work, for the first time (as far as we know), we exploit Reinforcement Learning (RL) to data-adaptively learn a pruning policy. Formulating token pruning as a sequential decision-making problem, we model it as a Markov Game and utilize Multi-Agent Proximal Policy Optimization (MAPPO) where each agent makes an individualized pruning decision for a single token. We also develop reward functions that enable simultaneous collaboration and competition of these agents to balance efficiency and accuracy. On the well-known ImageNet-1k dataset, our method improves the inference speed by up to 44% while incurring only a negligible accuracy drop of 0.4%. The source code is available at https://***/daashuai/rl4evit. © 2025, CC BY.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Rapid Coverage Control with Multi-agent systems Based on K-Means Algorithm 7

Rapid Coverage Control with Multi-agent Systems Based on K-M...

引用

7th International Conference on Information, Cybernetics, and Computational Social systems, ICCSS 2020

作者： Feng, YuZe Lu, Gang Bai, WenJie Zhao, JianHang Bai, YuMo Xu, Tao School of Mechanical Engineering Northwestern Polytechnical University Xi'an China College of Information Science and Engineering Hebei North University Zhang jiakou China School of Computer Science and Engineering South China University of Technology Guangzhou China Unmanned System Research Institute Northwestern Polytechnical University Xi'an China

ISBN: (纸本)9781728162461

This paper presents a strategy for covering particles in a specific region with multi-agent systems. Previous strategies that study on partitioning regions, have struggled to cover complex regions. The proposed approach in this paper, based on k-Means algorithm, can partition particles in some complex regions and calculate the locations of agents which have limited range of detection. And then the agents can cooperate to monitor the whole particles. Simulation results show that the particles in different regions can be covered well by multi-agent systems based on k-Means algorithm. © 2020 IEEE.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Convolutional Neural Network Based Career Recommender system for Pakistani engineering Students

Convolutional Neural Network Based Career Recommender System...

引用

International Conference on Innovative Computing (ICIC)

作者： Takreem Saeed Muhammad Sufian Mubashir Ali Attique Ur Rehman Department of Computer Engineering College of Electrical & Mechanical Engineering (CEME) National University of Sciences & Technology (NUST) Islamabad Pakistan Department of Software Engineering Lahore Garrison University (LGU) Lahore Pakistan School of System and Technology University of Management and Technology (UMT) Lahore Pakistan

ISBN: (数字)9781665400916

ISBN: (纸本)9781665400923

In recent years, Recommender systems are utilized in a variety of areas. One reason behind why we want a recommender system in current society is that an individual has many alternatives to use because of the pervasiveness of the Internet. A recommender system seeks to estimate and predict user content preference. Old recommender systems used State-of-the-art recommender algorithms like content based filtering to predict ratings. Career Recommender system provides engineering candidates the best possible available jobs relevant to their skills, qualification, etc. Four to six major engineering disciplines are covered in this recommender system. The proposed approach is tested using a career recommendation dataset which is collected from many students of different disciplines of various universities. A deep NLP based CNN model is used to predict the best jobs with maximum precision. 512 hidden layers are used to increase the performance of this system. Career recommendation takes care of the users and saves their cost and time spending on traditional job searching methods. Comparative study demonstrates that the proposed methodology of prediction of the best jobs achieves better results with an accuracy of 84% when matched with content based filtering technique where 81% accuracy is gained for content based career recommender system.

关键词： Costs engineering profession Filtering Neural networks Natural languages Predictive models Prediction algorithms

来源：评论

学校读者我要写书评

暂无评论

Object-oriented Map Exploration and Construction Based on Auxiliary Task Aided DRL

Object-oriented Map Exploration and Construction Based on Au...

引用

International Conference on Pattern Recognition

作者： Junzhe Xu Jianhua Zhang Shengyong Chen Honghai Liu College of Computer Science and Technology Zhejiang University of Technology Hangzhou China College of Computer Science and Technology Tianjin University of Technology Tianjin China State Key Laboratory of Robotics and System School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen China

Environment exploration by autonomous robots through deep reinforcement learning (DRL) based methods has attracted more and more attention. However, existing methods usually focus on robot navigation to single or multiple fixed goals, while ignoring the perception and construction of external environments. In this paper, we propose a novel environment exploration task based on DRL, which requires a robot fast and completely perceives all objects of interest, and reconstructs their poses in a global environment map, as much as the robot can do. To this end, we design an auxiliary task aided DRL model, which is integrated with the auxiliary object detection and 6-DoF pose estimation components. The outcome of auxiliary tasks can improve the learning speed and robustness of DRL, as well as the accuracy of object pose estimation. Comprehensive experimental results on the indoor simulation platform AI2-THOR have shown the effectiveness and robustness of our method.

关键词： Navigation Object oriented modeling Pose estimation Semantics Object detection Reinforcement learning Robustness

来源：评论

学校读者我要写书评

暂无评论

A systematic Review of Multi-Sensor Information Fusion for Equipment Fault Diagnosis

引用

IEEE Transactions on Instrumentation and Measurement 2025年

作者： Lin, Tantao Ren, Zhijun Zhu, Linbo Zhu, Yongsheng Feng, Ke Ding, Weiping Yan, Ke Beer, Michael Xi'an Jiaotong University Key Laboratory of Education Ministry for Modern Design & Rotor-Bearing System Xi'an710049 China Xi'an Jiaotong University Frontier Institute of Science and Technology Xi'an710049 China Xi'an Jiaotong University School of School of Chemical Engineering and Technology Xi'an710049 China Xi'an Jiaotong University School of Mechanical Engineering Xi'an710049 China Nantong University School of Artificial Intelligence and Computer Science Nantong226019 China City University of Macau Faculty of Data Science Macau999078 China Leibniz University Institute for Risk and Reliability Hannover30167 Germany University of Liverpool Department of Civil and Environmental Engineering LiverpoolL69 3BX United Kingdom Tongji University International Joint Research Center for Resilient Infrastructure International Joint Research Center for Engineering Reliability and Stochastic Mechanics Shanghai200092 China

In contrast to fault diagnosis relying solely on a single sensor, the method of multi-sensor information fusion for fault diagnosis (MSIFFD) broadens the spectrum of available information sources. It is renowned for its high accuracy and reliability, traits that have attracted growing attention within the research community and led to the generation of a substantial body of publications. However, there is currently a lack of a comprehensive and systematic review in this domain. Thereby, this review aims to thoroughly explore all research achievements in the field of MSIFFD. At the outset, an analysis is undertaken to delineate the fusion level of multi-sensor information, with a specific focus on its location and types of inputs and outputs within the information flow. Subsequently, an examination of the six primary fundamental operations for amalgamating multi-sensor information is undertaken to clarify the motivations and processes involved in information fusion across various operations. Following this, factors influencing fusion diagnostics and strategies aimed at improving their performance are examined to explore approaches for enhancing fusion accuracy. The subsequent sections delve into the analysis of sensor types and application scenarios, providing a reference guide for practical applications. Finally, this review outlines potential future challenges in the field of MSIFFD and provides a range of recommendations and possible solutions for consideration. © 1963-2012 IEEE.

关键词： Information fusion

来源：评论

学校读者我要写书评

暂无评论

Task-level authoring for remote robot teleoperation

arXiv

引用

arXiv 2021年

作者： Senft, Emmanuel Hagenow, Michael Welsh, Kevin Radwin, Robert Zinn, Michael Gleicher, Michael Mutlu, Bilge Department of Computer Sciences University of Wisconsin–Madison MadisonWI United States Department of Mechanical Engineering University of Wisconsin–Madison MadisonWI United States Department of Industrial System and Engineering University of Wisconsin–Madison MadisonWI United States

Remote teleoperation of robots can broaden the reach of domain specialists across a wide range of industries such as home maintenance, health care, light manufacturing, and construction. However, current direct control methods are impractical, and existing tools for programming robot remotely have focused on users with significant robotic experience. Extending robot remote programming to end users, i.e., users who are experts in a domain but novices in robotics, requires tools that balance the rich features necessary for complex teleoperation tasks with ease of use. The primary challenge to usability is that novice users are unable to specify complete and robust task plans to allow a robot to perform duties autonomously, particularly in highly variable environments. Our solution is to allow operators to specify shorter sequences of high-level commands, which we call task-level authoring, to create periods of variable robot autonomy. This approach allows inexperienced users to create robot behaviors in uncertain environments by interleaving exploration, specification of behaviors, and execution as separate steps. End users are able to break down the specification of tasks and adapt to the current needs of the interaction and environments, combining the reactivity of direct control to asynchronous operation. In this paper, we describe a prototype system contextualized in light manufacturing and its empirical validation in a user study where 18 participants with some programming experience were able to perform a variety of complex telemanipulation tasks with little training. Our results show that our approach allowed users to create flexible periods of autonomy and solve rich manipulation tasks. Furthermore, participants significantly preferred our system over comparative more direct interfaces, demonstrating the potential of our approach for enabling end users to effectively perform remote robot programming. Copyright © 2021, The Authors. All rights reserved.

关键词： Robot programming

来源：评论

学校读者我要写书评

暂无评论

Resource allocation network and PID control based on automated guided vehicles

引用

Academic Journal of Manufacturing engineering 2019年第2期17卷 137-144页

作者： Wang, Qizhi Wang, Xiaoxia Sun, Bingyang College of Mechanical Engineering and Automation Huaqiao University Key Laboratory of Process Monitoring and System Optimization for Mechanical and Electrical Equipment in Fujian Province University Huaqiao University Xiamen361021 China College of Computer Science and Technolog Huaqiao University Xiamen361021 China College of Mechanical Engineering and Automation Huaqiao University Xiamen361021 China

Automated Guided Vehicles (AGV) are used more and more in modern industrial production and logistics. In the traditional guiding method, the AGV mainly adopts the magnetic guiding mode, and the obtained deviation information is single, and the control effect is poor, so that the operation of the AGV is easy to cause the left and right shaking phenomenon, thereby causing the AGV to derail. In the actual factory floor operation, the traditional PID (Proportional Integral Derivative) controller control technology is difficult to meet the complex trajectory tracking requirements of AGV. Therefore, the research on the path recognition of AGV and its trajectory tracking control has extremely important application value. The core of AGV car motion control is to control servo drive motor to achieve speed tracking. In this paper, PID parameters are adjusted by RAN algorithm. The simulation results show that the method has high performance, good tracking ability and high approximation precision. © 2019 Editura Politechnica. All rights reserved.

关键词： Three term control systems

来源：评论

学校读者我要写书评

暂无评论

Tunable Non-Volatile Analog Resistive Memory and Its Application in AI

Tunable Non-Volatile Analog Resistive Memory and Its Applica...

引用

Midwest Symposium on Circuits and systems (MWSCAS)

作者： Heba Abunahla Yasmin Halawani Baker Mohammad Anas Alazzam System on Chip Center Electrical Engineering and Computer Science Department Khalifa University Abu Dhabi UAE System on Chip Center Mechanical Engineering Department Khalifa University Abu Dhabi UAE

ISBN: (数字)9781728180588

ISBN: (纸本)9781728180595

This paper presents a novel analog resistive random-access memory (RRAM), named NeuroMem, which consists of Au/GO/Au. The device's resistance can be tuned to any value within its R OFF to R ON range with high precision. The analog characteristic of NeuroMem mimics the memorization behavior of the brain, which makes it great asset for artificial neural network applications. In this work, NeuroMem-based crossbars are fabricated to hold the synaptic weights needed to perform Iris classification. The weight values are mapped to conductance states within NeuroMem R OFF to R ON range, and then written accurately on the actual devices. Unlike other RRAM-based hardware with limited conductance states, in this work no quantization is needed which enables efficient in memory-computing without scarifying accuracy. Furthermore, Neuromem device has been demonstrated in crossbars on flexible polymer substrate using standard photolithography process, which facilitates producing low cost flexible electronics. This work opens up great insights towards realizing RRAM-based computing at the edge.

关键词： Memristors Resistance Artificial intelligence Graphene Electrodes Switches Iris

来源：评论

学校读者我要写书评

暂无评论

Reversible Data Hiding Based on Pixel-Value-Ordering and Pixel Block Merging Strategy

引用

computers, Materials & Continua 2019年第6期59卷 925-941页

作者： Wengui Su Xiang Wang Yulong Shen College of Computer Science and Technology Xidian UniversityXi’an710071China Guangxi Key Laboratory of Manufacturing System&Advanced Manufacturing Technology College of Mechanical EngineeringGuangxi UniversityNanning530004China College of Telecommunications Engineering Xidian UniversityXi’an710071China

With the reversible data hiding method based on pixel-value-ordering,data are embedded through the modification of the maximum and minimum values of a block.A significant relationship exists between the embedding performance and the block *** pixel-value-ordering methods utilize pixel blocks with a fixed size to embed data;the smaller the pixel blocks,greater is the embedding ***,it tends to result in the deterioration of the quality of the marked ***,a novel reversible data hiding method is proposed by incorporating a block merging strategy into Li et al.’s pixel-value-ordering method,which realizes the dynamic control of block size by considering the image ***,the cover image is divided into non-overlapping 2×2 pixel ***,according to their complexity,similarity and thresholds,these blocks are employed for data embedding through the pixel-value-ordering method directly or after being emerged into 2×4,4×2,or 4×4 sized ***,smaller blocks can be used in the smooth region to create a high embedding capacity and larger blocks in the texture region to maintain a high peak signal-to-noise *** results prove that the proposed method is superior to the other three advanced *** achieves a high embedding capacity while maintaining low distortion and improves the embedding performance of the pixel-value-ordering algorithm.

关键词： Reversible data hiding pixel-value-ordering prediction error expansion dynamic block partition

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：