检索结果-内蒙古大学图书馆

SSRN 2022年

作者： Tong, Junlong Xie, Liping Yang, Wankou Zhang, Kanjian School of Automation Southeast University Nanjing 210096 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing 210096 China

Time series forecasting is crucial for many fields, such as disaster warning, weather prediction, and energy consumption. The Transformer-based models are considered to have revolutionized the field of sequence modeling. However, the complex temporal patterns of the time series hinder the model from mining reliable temporal dependencies. Furthermore, the autoregressive form of the Transformer introduces cumulative errors in the inference step. In this paper, we propose the probabilistic decomposition Transformer model that combines the Transformer with a conditional generative model, which provides hierarchical and interpretable probabilistic forecasts for intricate time series. The Transformer is employed to learn temporal patterns and implement primary probabilistic forecasts, while the conditional generative model is used to achieve non-autoregressive hierarchical probabilistic forecasts by introducing latent space feature representations. In addition, the conditional generative model reconstructs typical features of the series, such as seasonality and trend terms, from probability distributions in the latent space to enable complex pattern separation and provide interpretable forecasts. Extensive experiments on several datasets demonstrate the effectiveness and robustness of the proposed model, indicating that it compares favorably with the state of the art. © 2022, The Authors. All rights reserved.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

Transformer Based Multi-modal Fusion For Place Recognition with Self-attention Mechanism

Transformer Based Multi-modal Fusion For Place Recognition w...

引用

Chinese Automation Congress (CAC)

作者： Yan Pan Xiang Yu Zeqing Chang Fang Fang Bo Zhou Key Laboratory of Measurement and Control of Complex Systems of Engineering School of Automation Southeast University Ministry of Education Nanjing P. R. China

Autonomous driving still leaves a challenging task that how to apply the complementary information captured from different sensors, i.e. cameras and LiDAR, to handle place recognition task. In this paper, a brand new pipeline was designed to tackle the retrieval problem with a multi-modal late fusion. The light-weighted convolutional blocks is applied to encode features from images, followed by converting them into pseudo point clouds. Inspired by the progress of the Transformer, the network employs a residual point transformer module to extract feature vectors for 3D points and pseudo point clouds respectively. Finally, the two corresponding local descriptors are fused to get a robust fused global descriptor, which is capable of place recognition task after end-to-end training. Experimental results collected from sequence 00, 02, 05, 06 of KITTI dataset validate that fusing information from two modalities is able to enhance performance.

关键词： Point cloud compression Training Laser radar Three-dimensional displays Streaming media Transformers Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Phase-Shifting Coder: Predicting Accurate Orientation in Oriented Object Detection

arXiv

引用

arXiv 2022年

作者： Yu, Yi Da, Feipeng School of Automation Southeast University Nanjing China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Southeast University Nanjing China

With the vigorous development of computer vision, oriented object detection has gradually been featured. In this paper, a novel differentiable angle coder named phase-shifting coder (PSC) is proposed to accurately predict the orientation of objects, along with a dual-frequency version (PSCD). By mapping the rotational periodicity of different cycles into the phase of different frequencies, we provide a unified framework for various periodic fuzzy problems caused by rotational symmetry in oriented object detection. Upon such a framework, common problems in oriented object detection such as boundary discontinuity and square-like problems are elegantly solved in a unified form. Visual analysis and experiments on three datasets prove the effectiveness and the potentiality of our approach. When facing scenarios requiring high-quality bounding boxes, the proposed methods are expected to give a competitive performance. The codes are publicly available at https://***/open-mmlab/mmrotate. Copyright © 2022, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Parallax-Tolerant Image Stitching with Epipolar Displacement Field

arXiv

引用

arXiv 2023年

作者： Yu, Jian Da, Feipeng School of Automation Southeast University No.2 Sipailou Jiangsu Nanjing210096 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education No.2 Sipailou Jiangsu Nanjing210096 China

Image stitching with parallax is still a challenging task. Existing methods often struggle to maintain both the local and global structures of the image while reducing alignment artifacts and warping distortions. In this paper, we propose a novel approach that utilizes epipolar geometry to establish a warping technique based on the epipolar displacement field. Initially, the warping rule for pixels in the epipolar geometry is established through the infinite homography. Subsequently, the epipolar displacement field, which represents the sliding distance of the warped pixel along the epipolar line, is formulated by thin-plate splines based on the principle of local elastic deformation. The stitching result can be generated by inversely warping the pixels according to the epipolar displacement field. This method incorporates the epipolar constraints in the warping rule, which ensures high-quality alignment and maintains the projectivity of the panorama. Qualitative and quantitative comparative experiments demonstrate the competitiveness of the proposed method for stitching images with large parallax. Copyright © 2023, The Authors. All rights reserved.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Modified Deep Deterministic Policy Gradient Based Optimal Speed Tracking Strategy for PMSMs

Modified Deep Deterministic Policy Gradient Based Optimal Sp...

引用

Data Driven control and Learning systems (DDCLS)

作者： Zhong-Xin Fan Lintao Tang Shihua Li Rongjie Liu School of Automation Southeast University Nanjing P. R. China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing P. R. China Department of Statistics Florida State University Tallahassee FL USA

ISBN: (数字)9798350361674

ISBN: (纸本)9798350361681

Achieving optimal speed regulation for permanent magnet synchronous motors (PMSMs) remains a challenging task, particularly in selecting the most suitable controller to meet desired objectives. This paper considers the optimal speed tracking problem of PMSMs. We employ a reinforcement learning algorithm to obtain optimal controller with a linear form. The reinforcement learning approach incorporates a critic-actor network structure, which approximates the value function and optimal control, respectively. Specifically, a weight updating method for the critic network is introduced, where the update of the actor network is guided by the critic network. Through data collection and neural network training, dynamic acquisition of the optimal controller gain is achieved. In addition, a novel activation function is implemented in the network, leading to improved control performance concerning dataset size and reduced training time. At final, a real-world PMSM application is presented along with comparisons to demonstrate effectiveness. The desire idea of resulting optimal controllers is not only applicable to linear systems but also capable of approximating arbitrary nonlinear optimal controllers, thus offering practical engineering versatility.

关键词： Training Linear systems Learning systems Neural networks Optimal control Reinforcement learning Data collection

来源：评论

学校读者我要写书评

暂无评论

TransTracking for UAV: An Autonomous Real-time Target Tracking System for UAV via Transformer Tracking

TransTracking for UAV: An Autonomous Real-time Target Tracki...

引用

2021 International Conference on Intelligent Technology and Embedded systems, ICITES 2021

作者： Sun, Xiaolou Xie, Fei Wang, Qi Yao, Yuncong Wang, Hao Wang, Wei Yang, Wankou School of Automation Southeast University Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing210096 China Nanjing University of Information Science and Technology School of Automation Nanjing211800 China

ISBN: (纸本)9781665427555

Most of the existing Siamese-type trackers usually adopt pre-defined anchor boxes or anchor-free schemes to accurately estimate the bounding box of targets. Unfortunately, they suffer from complicated hand-designed components and tedious post-processings. It is not easy to adjust parameters for unique scenes in real applications. So, we propose a new scheme by formulating visual tracking as a direct set prediction problem to alleviate this issue. The main component is a transformer attached to the Siamese-type feature extraction networks. Thus, our new framework can be summarized as Siamese Network with Transformers (SiamTFR). With a fixed small set of learned object queries, we force the final set of predictions via bipartite matching, significantly reducing hyper-parameters associated with the candidate boxes. Due to the unique predictions of this framework, we significantly ease the heavy burden of hyper-parameters search of post-processings in visual tracking. Extensive experiments on visual tracking benchmarks, including GOT-10K, demonstrate that SiamTFR achieves competitive performance and runs at 50 FPS. Specifically, SiamTFR outperforms leading anchor-based tracker SiamRPN++ in the GOT-10K benchmark, confirming its effectiveness and efficiency. Furthermore, SiamTFR is deployed on the embedded device in which the algorithm can be run at 30FPS or 54FPS with TensorRT meeting the real-time requirements. In addition, we design the complete tracking system demo that can work in the real road to narrow the gap between the academic models and industrial deployments. © 2021 IEEE.

关键词： Unmanned aerial vehicles (UAV)

来源：评论

学校读者我要写书评

暂无评论

Wind Speed Data Repairing Method Based on Bidirectional Prediction

Wind Speed Data Repairing Method Based on Bidirectional Pred...

引用

2021 IEEE International Conference on Power Electronics, Computer Applications, ICPECA 2021

作者： Shen, Xincheng Qu, Yi Huang, Shaoxiong Li, Zhi Zhang, Kaifeng Southeast University Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing China State Grid Anhui Electric Power Co. Ltd Department of Dispatch Center Hefei China

ISBN: (纸本)9781728190037

In order to repair the lost data in distributed wind power system, this paper puts forward a wind speed data repairing model based on a new bidirectional prediction method. This model consists of two one-way prediction models. In each prediction model, the original wind speed data are decomposed into several intrinsic mode functions (IMFs) and a residue signal by ensemble empirical mode decomposition (EEMD) method. Then the Savitzky-Golay (SG) filter is used to reduce noise for high-frequency IMFs. Next the long short-term memory (LSTM) model and autoregressive integrated moving average (ARIMA) model are combined to predict low-frequency IMFs and the noise reduction results respectively. At the end, all those forecast results are added and form a one-way result. By weighted average of two one -way results, the repairing result is calculated. The experimental results from multiple prediction cases show that this method can get more accurate results. © 2021 IEEE.

关键词： Empirical mode decomposition

来源：评论

学校读者我要写书评

暂无评论

A Semi-Supervised Railway Foreign Object Detection Method Based on GAN 21

A Semi-Supervised Railway Foreign Object Detection Method Ba...

引用

5th International Conference on Computer Science and Application engineering, CSAE 2021

作者： Chen, Yanqi Tong, Shuzhen Lu, Xiaobo Wei, Yun School of Automation Southeast University Nanjing China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Nanjing China Beijing Mass Transit Railway Operation Corporation Limited Beijing China

ISBN: (纸本)9781450389853

The rapid development of deep learning provides new technical means for railway foreign object detection. However, in practical applications, the datasets of railways with foreign objects are scarce. In order to solve this problem, by improving the loss function and anomaly image evaluation standard, this paper proposes a newsemisupervised anomaly detection method based on GAN (Generative Adversarial Networks). Experiments show that our method can achieve railway foreign object detection without anomaly prior knowledge. Regarding anomaly recognition, a 0.058 AUC (Area Under Curve) and a 6% classification accuracy relative improvement for the railway dataset used in this paper are obtained.. © 2021 Association for Computing Machinery. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Efficient Wi-Fi AP Localization through Channel Feature Fusion and Anomaly Detection

Efficient Wi-Fi AP Localization through Channel Feature Fusi...

引用

IEEE Conference on Wireless Communications and Networking

作者： Yan Li Jie Yang Shang-Ling Shih Wan-Ting Shih Chao-Kai Wen Shi Jin National Mobile Communications Research Laboratory Southeast University Nanjing China Frontiers Science Center for Mobile Information Communication and Security Nanjing China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education Southeast University Nanjing China Institute of Communications Engineering National Sun Yat-sen University Kaohsiung Taiwan

ISBN: (数字)9798350303582

ISBN: (纸本)9798350303599

Wi-Fi access point (AP) and IoT device localization are essential for smart home functionalities, including indoor localization and privacy protection. Yet, complex multipath channels in indoor settings often hinder precise localization. To overcome this, we introduce an Artificial Intelligence (AI) technique that amalgamates channel state information from proximate trajectory points, thus elevating the accuracy of line of sight (LoS) angle of arrival (AoA) estimation. Our methodology initiates with an AI-based anomaly detection system to eliminate questionable measurements. Thereafter, our AI-optimized LoS-AoA network proficiently identifies the primary LoS path from the several multipaths detected by the multipath estimation process and autonomously fine-tunes the LoS-AoA estimation. Using simulations in an indoor office environment with Wireless Insite, our results reveal that our approach considerably improves LoS-AoA estimations, even under challenging indoor scenarios. Notably, our technique enhanced AP positioning accuracy in 68% of instances, reducing a 2-meter error to 0.6 meters, and in 95% of instances, cutting down a 10-meter error to 2 meters when measured against top benchmarks.

关键词： Location awareness Meters Wireless communication Accuracy Estimation Smart homes Trajectory

来源：评论

学校读者我要写书评

暂无评论

Study on Self-Organizing Aggregation Method for Massive Flexible Resources and Its Simulation

Study on Self-Organizing Aggregation Method for Massive Flex...

引用

Advanced Electrical and Energy systems (AEES), International Conference on

作者： Changwei Yin Wenlu Ji Ting Ye Tingfeng Mao Kaifeng Zhang Ministry of Education Key Laboratory of Measurement and Control of Complex Systems of Engineering Southeast University Nanjing China Nanjing Power Supply Company of State Grid Jiangsu Electric Power Co. Ltd. Nanjing China

Distributed flexible resources are developing in the trend of sea quantization. In this situation, the existing resource aggregation methods are facing bottlenecks, and it is difficult to fully exploit the regulation potential of massive flexible resources. In order to achieve efficient and flexible aggregation in the scenario of massive resources, this paper proposes a self-organizing aggregation method for flexible resources based on the social network and simulates the aggregation process under the proposed method based on the multi-agent simulation platform NetLogo. Firstly, this paper outlines the idea of the self-organizing aggregation method for flexible resources, which projects flexible resources from the physical space to the social network space, and realizes flexible aggregation through the interactive attraction between resources in the social network. Secondly, this paper establishes a simulation model for flexible resource self-organizing aggregation based on the NetLogo platform. The model defines four types of agents involved in the self-organizing aggregation process and provides detailed specifications for their attributes and interaction rules. Finally, simulation experiments are designed to verify the rationality of the self-organizing aggregation method and simulation model proposed in this paper. The simulation results show that, following the idea of self-organizing aggregation, massive flexible resources can autonomously and flexibly aggregate through the interaction attraction in the social network due to the incentive of task gain, demonstrating the effectiveness and rationality of the self-organizing aggregation method and the simulation model proposed in this paper.

关键词：

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：