A Study on Reducing Traffic Congestion in the Roadside Unit for Autonomous Vehicles Using BSM and PVD

World Electric Vehicle Journal, Journal Year: 2024, Volume and Issue: 15(3), P. 117 - 117

Published: March 18, 2024

With the rapid advancement of autonomous vehicles reshaping urban transportation, importance innovative traffic management solutions has escalated. This research addresses these challenges through deployment roadside units (RSUs), aimed at enhancing flow and safety within driving era. Our research, conducted in diverse road settings such as straight circle roads, delves into RSUs’ capacity to diminish density alleviate congestion. Employing vehicle-to-infrastructure communication, we can scrutinize its essential role navigating vehicles, incorporating basic messages (BSMs) probe vehicle data (PVD) accurately monitor presence status. paper presupposes connectivity all contemplating integration on-board or diagnostics legacy extend connectivity, albeit this aspect falls beyond work’s current ambit. detailed experiments on two types roads demonstrate that behavior is significantly impacted when reaches critical thresholds 3.57% 34.41% roads. However, it important note identified threshold values are not absolute. In our experiments, represent points which one begins impact more vehicles. At levels, propose RSUs intervene mitigate issues by implementing measures prohibiting lane changes restricting entry circles. We a new message set PVD for RSUs: balance. Using message, negotiate between approach underscores capability actively manage prevent congestion, highlighting their maintaining optimal conditions safety.

Language: Английский

---

Real-Time Analysis of Industrial Data Using the Unsupervised Hierarchical Density-Based Spatial Clustering of Applications with Noise Method in Monitoring the Welding Process in a Robotic Cell

Information, Journal Year: 2025, Volume and Issue: 16(2), P. 79 - 79

Published: Jan. 22, 2025

The application of modern machine learning methods in industrial settings is a relatively new challenge and remains the early stages development. Current computational power enables processing vast numbers production parameters real time. This article presents practical analysis welding process robotic cell using unsupervised HDBSCAN algorithm, highlighting its advantages over classical k-means algorithm. paper also addresses problem predicting monitoring undesirable situations proposes use real-time graphical representation noisy data as particularly effective solution for managing such issues.

Language: Английский

---

Anomaly Detection in Connected and Autonomous Vehicle Trajectories Using LSTM Autoencoder and Gaussian Mixture Model

Electronics, Journal Year: 2024, Volume and Issue: 13(7), P. 1251 - 1251

Published: March 28, 2024

Connected and Autonomous Vehicles (CAVs) technology has the potential to transform transportation system. Although these new technologies have many advantages, implementation raises significant concerns regarding safety, security, privacy. Anomalies in sensor data caused by errors or cyberattacks can cause severe accidents. To address issue, this study proposed an innovative anomaly detection algorithm, namely LSTM Autoencoder with Gaussian Mixture Model (LAGMM). This model supports anomalous CAV trajectory real-time leveraging communication capabilities of sensors. The is applied generate low-rank representations reconstruct for each input point, while (GMM) employed its strength density estimation. was jointly optimized GMM simultaneously. utilizes realistic from a platooning experiment conducted Cooperative Automated Research Mobility Applications (CARMAs). findings indicate that LAGMM approach enhances accuracy 3% precision 6.4% compared existing state-of-the-art methods, suggesting improvement field.

Language: Английский

---

Energy-Efficient Anomaly Detection and Chaoticity in Electric Vehicle Driving Behavior

Sensors, Journal Year: 2024, Volume and Issue: 24(17), P. 5628 - 5628

Published: Aug. 30, 2024

Detection of abnormal situations in mobile systems not only provides predictions about risky but also has the potential to increase energy efficiency. In this study, two real-world drives a battery electric vehicle and unsupervised hybrid anomaly detection approaches were developed. The performances models created with combination Long Short-Term Memory (LSTM)-Autoencoder, Local Outlier Factor (LOF), Mahalanobis distance evaluated silhouette score, Davies–Bouldin index, Calinski–Harabasz recovery rates determined. Two driving datasets terms chaotic aspects using Lyapunov exponent, Kolmogorov–Sinai entropy, fractal dimension metrics. developed are superior sub-methods detection. Hybrid Model-2 had 2.92% more successful results compared Model-1. saving, Model-1 provided 31.26% superiority, while 31.48%. It was observed that there is close relationship between chaoticity. literature where cyber security visual sources dominate detection, strategy efficiency-based analysis from data obtained without additional sensor data.

Language: Английский

---

Single and Mixed Sensory Anomaly Detection in Connected and Automated Vehicle Sensor Networks

Electronics, Journal Year: 2024, Volume and Issue: 13(10), P. 1885 - 1885

Published: May 11, 2024

Connected and automated vehicles (CAVs), integrated with sensors, cameras, communication networks, are transforming the transportation industry providing new opportunities for consumers to enjoy personalized seamless experiences. The fast proliferation of connected on road growing trend autonomous driving create vast amounts data that need be analyzed in real time. Anomaly detection CAVs refers identifying any unusual or unforeseen behavior generated by vehicles’ various sensors components. aims identify might indicate a problem malfunction vehicle. To detect anomalies efficiently, method must deal noisy data, missing dynamic frequency low- high-magnitude it accurate enough sensor streaming environment. Therefore, this paper proposes efficient hard-voting-based technique named FT-HV, comprising three fine-tuned machine learning algorithms classify anomaly single mixed sensory datasets. In experiments, we evaluate our approach benchmark Sensor dataset contains from vehicle at low high magnitudes. Further, types challenging identify. results reveal proposed outperforms existing solutions detecting magnitudes all settings. Furthermore, research is envisioned help early efficiently promote safer more resilient CAVs.

Language: Английский

---

Network Information Security Monitoring Under Artificial Intelligence Environment

International Journal of Information Security and Privacy, Journal Year: 2024, Volume and Issue: 18(1), P. 1 - 25

Published: June 6, 2024

At present, network attack means emerge in endlessly. The detection technology of must be constantly updated and developed. Based on this, the two stages (feature selection traffic classification) are discussed. improved bat algorithm (O-BA) random forest (O-RF) proposed for optimization. Moreover, NIS system is designed based Agent concept. Finally, simulation experiment carried out real data platform. results showed that precision, accuracy, recall, F1 score O-BA significantly higher than those references [17], [18], [19], [20], while false positive rate opposite (P < 0.05). O-RF Apriori, ID3, SVM, NSA, algorithm, lower

Language: Английский

---

A Novel Hybrid Model (EMD-TI-LSTM) for Enhanced Financial Forecasting with Machine Learning

Mathematics, Journal Year: 2024, Volume and Issue: 12(17), P. 2794 - 2794

Published: Sept. 9, 2024

Financial forecasting involves predicting the future financial states and performance of companies investors. Recent technological advancements have demonstrated that machine learning-based models can outperform traditional techniques. In particular, hybrid approaches integrate diverse methods to leverage their strengths yielded superior results in prediction. This study introduces a novel model, entitled EMD-TI-LSTM, consisting empirical mode decomposition (EMD), technical indicators (TI), long short-term memory (LSTM). The proposed model delivered more accurate predictions than those generated by conventional LSTM approach on same well-known datasets, achieving average enhancements 39.56%, 36.86%, 39.90% based MAPE, RMSE, MAE metrics, respectively. Furthermore, show has lower MAPE rate 42.91% compared its state-of-the-art counterparts. These findings highlight potential mathematical innovations advance field forecasting.

Language: Английский

---

Machine Learning-Driven Calibration of Traffic Models Based on a Real-Time Video Analysis

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(11), P. 4864 - 4864

Published: June 4, 2024

Accurate traffic simulation models play a crucial role in developing intelligent transport systems that offer timely information to users and efficient management. However, calibrating these represent real-world conditions accurately poses significant challenge due the dynamic nature of flow limitations traditional calibration methods. This article introduces machine learning-based approach calibrate macroscopic using real-time video stream data. The proposed method for creating model has significantly improved statistical correspondence between generated vehicle characteristics real data about cars on simulated road section. increased from 37% 73%. Machine learning trained tested show accuracy rates. Mean absolute error, mean square percentage error decreased by more than two orders magnitude. coefficient determination also increased, approaching 1. eliminates need deploy wireless sensor networks, which can reduce cost implementing systems.

Language: Английский

---