检索结果-内蒙古大学图书馆

Resource allocation in UAV assisted air ground intelligent inspection system

学校读者我要写书评

暂无评论

Cognitive Robotics 2022年 2卷 1-12页

作者： Zhang, Zhuoya Xu, Fei Qin, Zengshi Xie, Yue School of ordnance science and technology Xi'an Technological University Xi'An China Engineering Laboratory The State and Provincial Joint Engineering Lab. of Advanced Network Monitoring and Control Xi'An China School of Computer Science and Engineering Xi'an Technological University Xi'An China

With the progress of power grid technology and intelligent technology, intelligent inspection robot (IR) came into being and are expected to become the main force of substation inspection in the future. Among them, mobile edge computing provides a promising architecture to meet the explosive growth of communication and computing needs of inspection robot. Inspection robot can transmit the collected High Definition (HD) video to adjacent edge servers for data processing and state research and judgment. However, the communication constraints of long-distance transmission, high reliability and low delay pose challenges to task offloading optimization. Therefore, this paper introduced Unmanned Aerial Vehicle (UAV) and established UAV assisted mobile edge computing system. UAV assisted and mobile edge computing are combined to form edge computing nodes. In this way, it provided communication and computing services to the IR for fast data processing. Specifically, in order to optimize the system energy consumption, a resource allocation strategy based on genetic algorithm is proposed. By optimizing the offloading decision and computing resource allocation of the IRs, the computing task of the IRs are offloaded to an energy-efficient UAV. The experimental results show that the resource allocation strategy based on genetic algorithm can effectively reduce the energy consumption and cost of UAVs and IRs, and effectively realize the reasonable allocation of resources. The results verify the effectiveness and reliability of the algorithm in the real scene. © 2021

关键词： Resource allocation

Wave Energy Prediction Based on Hybrid Frameworks Integrated with Vmd Quadratic Decomposition

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Wang, Jianhui Yang, Ruirong Cui, Zhendong School of Computer and Control Engineering Yantai University Yantai264000 China

Precise prediction of wave energy enables strategic planning for sustainable maritime transportation and optimal siting of power plants, thereby facilitating the transition towards greener energy solutions. However, wave energy exhibits high volatility with time, and intense energy output at specific moments. In this paper, two hybrid frameworks for wave energy prediction combined variational modal decomposition (VMD) were proposed, which integrating with sample entropy (SE), convolutional neural network (CNN), bidirectional long short-term memory (BiLSTM) network, and attention mechanism. For data preprocessing, raw data was decomposition-reconstruction-decomposition using VMD and SE to reduce data complexity and shorten runtime. For prediction, a CNN-BiLSTM (CBLA) hybrid model based on the attention mechanism was used to predict the processed data. Wave height, wave period, and other factors from three buoy stations in the North Pacific and Gulf of Mexico were used for the experiments. The proposed VMD-SE-VMD-CBLA framework in this paper enables the average wave energy prediction metrics MAPE, R2, MAE, and RMSE to reach 3.7309%, 99.72%, 0.4528kw/m, and 0.9956kw/m at the three stations. The results show that the proposed framework has higher accuracy and improved peak wave energy prediction compared to other benchmark models. © 2024, The Authors. All rights reserved.

关键词： Forecasting

Jointly Stochastic Fully Symmetric Interpolatory Rules and Local Approximation for Scalable Gaussian Process Regression

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Zhang, Hongli Liu, Jinglei School of Computer and Control Engineering Yantai University Yantai264005 China

The main application bottlenecks of large-scale Gaussian process regression lie in the following three points: 1) solving the inverse matrix of n training points results in higher O(n^3) time complexity;2) widely used local approximation methods may have inconsistent prediction result;3) many aggregation methods are not able to effectively differentiate the importance of the experts, which leads to inaccurate prediction results. In order to solve the above problems, this paper proposes a method that combines third-degree stochastic fully symmetric interpolatory rules and local approximation with Tsallis mutual information (TDSFSIRLA). First, TDSFSIRLA introduces a three-degree deterministic fully symmetric interpolatory rules to generate an adaptive dimensional feature mapping and approximate the Gaussian kernel. It not only greatly reduces the number of required orthogonal nodes, but also maintains a higher approximation accuracy. Second, we use the Generalized Robust Bayesian Committee Machine (GRBCM) to approximate local experts consistently. Finally, to address the problem of uneven distribution of expert weights, we introduce Tsallis mutual information to assign different weights for local experts. Experimental results on synthetic and seven real datasets verify that our method can greatly reduce the time complexity and improve the prediction accuracy. © 2024, The Authors. All rights reserved.

关键词： Forecasting

A Hybrid Deep Learning Framework for Accurate Prediction of Significant Wave Height in the North Pacific Ocean

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Wang, Jianhui Cui, Zhendong School of Computer and Control Engineering Yantai University Yantai264000 China

Significant wave height (WVHT) is one of the important parameters applied in the field of ocean engineering. Accurate predictions of WVHT can help improve wave energy conversion efficiency, coastal facility management, and disaster prevention and mitigation. However, extracting features from complex ocean parameters to accurately predict wave heights remains a challenge. A hybrid deep learning framework is proposed for precise prediction of WVHT, which combining the complementary strengths of deep learning architectures and gradient boosting methods. Incorporating an attention mechanism (AM), a convolutional neural network (CNN), and a bidirectional long and short-term memory network (BiLSTM), the framework is designed to deeply learn the information contained in historical data. Furthermore, the extreme gradient boosting (XGBoost) algorithm is integrated using a multi-model fusion strategy based on weighted ensemble techniques. First, input features are selected using Pearson's correlation coefficient and XGBoost feature importance score. The feature set is then fed into AM-based CNN-BiLSTM (CBLA) and XGBoost for training to predict WVHT. Finally, the predictions of the above models are fused to produce the final predictions. We predict the WVHT at three stations in the North Pacific at 1, 2, 4, and 6 hours in advance. The results show that the prediction performance and robustness of hybrid framework outperforms that of CBLA and XGBoost, and significantly outperforms that of other benchmark models. © 2024, The Authors. All rights reserved.

关键词： Forecasting

Fault Detection Method of Maintenance Equipment Based on ECAT 5

学校读者我要写书评

暂无评论

Fault Detection Method of Maintenance Equipment Based on ECA...

5th International Conference on Smart Grid and Energy engineering, SGEE 2024

作者： Qin, Shuo Zhang, Yimei Zhang, Yinghao Wang, Hongchao Chang, Bo Feng, Hao State Grid Beijing Electric Power Company Changping Power Supply Company Beijing China North China Electric Power University School of Control and Computer Engineering Beijing China

ISBN: (纸本)9798331530686

To improve the fault detection performance of power system operation and maintenance equipment, this paper studies the ECAT model through the method of integrating Empirical Mode Decomposition (EMD), Convolutional Neural Networks (CNN), and Attention mechanisms. The results show that: the ECAT model significantly outperforms traditional methods in fault detection accuracy and robustness on multiple public datasets, with an average accuracy improvement of about 10%;the model parameters are globally optimized using the Zebra Optimization Algorithm (ZOA), further enhancing detection performance;and the model effectively captures critical information and long-range dependencies. It can be seen that ECAT provides an efficient and accurate solution for intelligent fault detection of power system operation and maintenance equipment, demonstrating good application prospects. © 2024 IEEE.

关键词： Empirical mode decomposition

Research on the Probability Interval and Ramping Prediction Method of Wind Turbine Power

学校读者我要写书评

暂无评论

Research on the Probability Interval and Ramping Prediction ...

第43届中国控制会议

作者： Ketao Dong Yali Zhao Qiankun Hu Peng Guo School of Control and Computer Engineering North China Electric Power University

ISBN: (数字)9789887581581

ISBN: (纸本)9798350366907

The wind power ramp event refers to the large fluctuation of wind power caused by the sudden increase or decrease of wind power in a short time interval,which affects the safe and stable operation of the power grid *** article proposes a wind power prediction method based on the CNN-LSTM model,and the prediction interval of wind power is obtained by using a non-parametric kernel density estimation *** using interval prediction lower limit data combined with a detection method based on statistical power fluctuations and a Swing Door Trending algorithm proposed to identify wind power ramp events,wind power ramp event prediction is ultimately achieved,and so as to make countermeasures in *** results show that the proposed method is more beneficial to the prediction of wind power ramp events.

关键词： Wind power ramp event prediction wind power prediction non-parametric kernel density estimation

Medical Named Entity Recognition Based on Multiple Features and Dynamic Fusion

学校读者我要写书评

暂无评论

Medical Named Entity Recognition Based on Multiple Features ...

Data Science and Network Security (ICDSNS), International Conference on

作者： Yi Ren Wei Mao School of Computer and Control Engineering Shenyang Jianzhu University Shenyang China

ISBN: (数字)9798350373110

ISBN: (纸本)9798350373127

To Address the challenges of unclear entity delineations and insufficient utilization of Semantic data, This study introduces a novel fusion approach leveraging multiple features for dynamic integration. To enrich the semantic representation of text, Textual model's embedding layer incorporates diverse techniques. Firstly, convolutional neural networks are used to implement font embedding, enriching the character representation of text through Chinese character fonts. Secondly, SoftLexicon is used to fuse word information from the dictionary and enhance entity boundary information. To achieve multi feature embedding, the word vector's semantic information is modeled using McBERT. In the feature extraction layer, long-distance inter-character semantic information is obtained by IDCNN, whereas contextual semantic information is obtained via BiLSTM. At the conclusion, the utilization of the dynamic fusion methodology is employed to accomplish the task of recognizing named entities, leveraging the conditional random field model. The model demonstrated an F1 score of 88.96% on the Chinese medical information evaluation dataset, surpassing the BERT BiLSTM CRF model by 3.81%, thereby validating its effectiveness.

关键词： Technological innovation Dictionaries Text recognition Semantics Named entity recognition Network security Feature extraction

Multiobjective Multiverse Optimizer for Multirobotic U-Shaped Disassembly Line Balancing Problems

学校读者我要写书评

暂无评论

IEEE Transactions on Artificial Intelligence

IEEE Transactions on Artificial Intelligence 2024年第2期5卷 882-894页

作者： Qin, Shujin Zhang, Shancheng Wang, Jiacun Liu, Shixin Guo, Xiwang Qi, Liang School of Economics and Management Shangqiu Normal University Shangqiu476000 China School of Information and Control Engineering Liaoning Petrochemical University Fushun113001 China Monmouth University West Long Branch NJ07764 United States College of Information Science and Engineering Northeastern University Shenyang110819 China Department of Computer Science and Technology Shandong University of Science and Technology Qingdao266590 China

The development of technology accelerates the upgrade of products, which results in a significant number of obsolete products. This research aims to solve the multirobotic multiproduct U-shaped disassembly line balancing problem (M2UDP), in which different products are disassembled on a U-shaped model in a preset cycle time and assigned tasks to robots in each workstation reasonably. A linear mixed-integer model is established to maximize disassembly profit and minimize carbon emissions. An improved multiobjective multiverse optimizer (IMMO) that utilizes the sigmoid activation function in neural networks is proposed to find the optimal plan for the model. The improved algorithm is verified via a set of real-life instances and compared with three classical multiobjective optimization algorithms. The experimental results show that the proposed IMMO performs better than those peers in solving the M2UDP problems. © 2023 IEEE.

关键词： Profitability

Automation Detection, Segmentation, and Classification for Ocean Reinforced Concrete Corrosion Based on Artificial Intelligence

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Cui, Zhendong Jia, Zhitong Zhao, Yang Fu, Mingjiao School of Computer and Control Engineering Yantai University Yantai264000 China

Based on artificial intelligence technology, it is of great significance to automatically identify and determine the degree of corrosion damage for Ocean reinforced ***, an enhanced and comprehensive non-destructive testing framework was proposed for the automatic detection of corrosion in ocean concrete and the evaluation of its corrosion ***, a dataset containing 3,700 images of offshore ocean concrete corrosion has been established, which captured in various geographical locations, weather conditions, lighting situations, and engineering ***, a Faster-DeepLabV3+ model for damage segmentation and grade classification was present. An Efficient Channel Attention structure was introduced at the front of the inverted residual module of MobileNetV2 to reduce the model parameters and computation while maintaining high-accuracy segmentation results;a stick pooling layer was introduced in Atrous Spatial Pyramid Pooling to improve the model's ability to perceive damage at different scales;a combination of loss functions with Dice loss function and focal loss function was used to enhance the model's capacity to detect outlines and finer ***, different corrosion levels were identified to categorize cracks, spalling, and rebar corrosion in marine concrete into three conditions, and accurate delineation of corroded areas and rebar rusting was successfully ***, experiments are carried out on the constructed data set, and by comparing the intersection merging ratio, average pixel accuracy, pixel accuracy and frames per second, the proposed model is proved to be accurate and efficient. © 2024, The Authors. All rights reserved.

关键词： Nondestructive examination