Bearing defects occur under harsh working conditions and due to poor maintenance. Early diagnosis of bearing failures is crucial for the safe operation of rotating machinery. Modern approaches involving the classifica...
详细信息
This paper presents a pioneering data-centric approach for model identification of inverter-based resources (IBRs) in smart grids, which includes renewable energy systems and energy storage technologies. Unlike tradit...
详细信息
ISBN:
(纸本)9798350361612;9798350361629
This paper presents a pioneering data-centric approach for model identification of inverter-based resources (IBRs) in smart grids, which includes renewable energy systems and energy storage technologies. Unlike traditional methods that depend on fixed models and extensive system identification tools, our approach leverages emerging systems behavioral theories and combines them with singular value decomposition (SVD) to efficiently identify IBR models from data. The SVD's ability to reduce the dimension of collected data is instrumental in capturing critical dynamic features from minimal data inputs. By applying the principle of persistence of excitation and organizing input/output data into a Hankel matrix form, we derive a robust, model-free representation of IBR dynamics that requires significantly less data than conventional machine learning methods. The effectiveness of our approach is validated through comprehensive time-domain simulations, demonstrating its potential for model-free IBR control applications.
Urban traffic state estimation aims at providing accurate and reliable information about traffic flow characteristics, which can be used for urban traffic management. Traditional estimation approaches mainly use loop ...
详细信息
ISBN:
(纸本)9798350399462
Urban traffic state estimation aims at providing accurate and reliable information about traffic flow characteristics, which can be used for urban traffic management. Traditional estimation approaches mainly use loop detectors and/or floating car data, which are labour- and cost-intensive. With the ongoing technological development in autonomous driving, more and more research focuses on the use of onboard sensor data. In this paper, a novel concept for the collection and utilization of traffic data is presented: Autonomous Vehicles as a Sensor. The proposed concept capitalizes on the advanced sensor technologies deployed in autonomous vehicles, particularly those employed in Mobility-on-Demand services, to collect link-level traffic states. These microscopic traffic states are further utilized by e.g., Mobility-on-Demand operators to estimate network-level traffic states. A first proof of concept was examined through a case study in a microscopic traffic simulation with a grid network and generic demand. The results demonstrate that both, moving and parked autonomous vehicles, can effectively contribute to the estimation of the macroscopic fundamental diagram. When the results from both are combined, the resulting estimation yields the most accurate fit compared to the ground truth. These findings underline the potential of the 'Autonomous Vehicles as a Sensor' concept for accurate and reliable traffic state estimation.
This article presents a data interface for maritime simulators to facilitate the development of ship Automatic Navigational systems (ANS) at different autonomy levels. The core architecture of this data interface cons...
详细信息
The integration of healthcare data storage and management within cloud systems has revolutionized accessibility and administration in the medical field. This paper, proposes a security framework for medical data over ...
详细信息
Training of Machine Learning (ML) models requires huge amounts of data. Usually, in the training of sophisticated models, data-sets can be very computationally intensive on resources. In order to reduce the computatio...
详细信息
The emergence of many-core processors presents significant opportunities for large-scale multithreading. Exploiting these intensive computing resources poses an urgent challenge for data processing systems. Although c...
详细信息
Terminal devices are critical in establishing efficient, reliable, and secure computer network systems. Anomaly detection, a key technique in computer security, is widely researched for its adaptability and ability to...
详细信息
ISBN:
(纸本)9798350350920
Terminal devices are critical in establishing efficient, reliable, and secure computer network systems. Anomaly detection, a key technique in computer security, is widely researched for its adaptability and ability to recognize novel attacks, despite facing issues such as low efficiency. To address the monitoring and identification of anomalous behaviors by users during data access at the terminal level, a comprehensive assessment incorporating data content, methods of operation, and user behavior is essential. Therefore, a terminal-level data access anomaly detection system utilizing the TL-IDA algorithm is proposed. This system employs a sliding window technique to enhance its performance. The primary aim is to increase data security and prevent unauthorized access, providing guidance for the prompt adoption of appropriate security measures to ensure the protection of data.
This article studies the data-driven design of parity space based replay attack detection method for cyber-physical systems. By utilizing the input and output data, the proposed data-driven detection can uncover steal...
详细信息
We address the challenges of the semi-supervised LiDAR segmentation (SSLS) problem, particularly in low-budget scenarios. The two main issues in low-budget SSLS are the poor-quality pseudo-labels for unlabeled data, a...
详细信息
ISBN:
(纸本)9798350377712;9798350377705
We address the challenges of the semi-supervised LiDAR segmentation (SSLS) problem, particularly in low-budget scenarios. The two main issues in low-budget SSLS are the poor-quality pseudo-labels for unlabeled data, and the performance drops due to the significant imbalance between ground-truth and pseudo-labels. This imbalance leads to a vicious training cycle. To overcome these challenges, we leverage the spatio-temporal prior by recognizing the substantial overlap between temporally adjacent LiDAR scans. We propose a proximity-based label estimation, which generates highly accurate pseudo-labels for unlabeled data by utilizing semantic consistency with adjacent labeled data. Additionally, we enhance this method by progressively expanding the pseudo-labels from the nearest unlabeled scans, which helps significantly reduce errors linked to dynamic classes. Additionally, we employ a dual-branch structure to mitigate performance degradation caused by data imbalance. Experimental results demonstrate remarkable performance in low-budget settings (i.e., <= 5%) and meaningful improvements in normal budget settings (i.e., 5 - 50%). Finally, our method has achieved new state-of-the-art results on SemanticKITTI and nuScenes in semi-supervised LiDAR segmentation. With only 5% labeled data, it offers competitive results against fully-supervised counterparts. Moreover, it surpasses the performance of the previous state-of-the-art at 100% labeled data (75.2%) using only 20% of labeled data (76.0%) on nuScenes. The code is available on https://***/halbielee/PLE.
暂无评论