A Hybrid Deep Reinforcement Learning and Optimal Control Architecture for Autonomous Highway Driving DOI Creative Commons
Nicola Albarella, Dario Giuseppe Lui, Alberto Petrillo

et al.

Energies, Journal Year: 2023, Volume and Issue: 16(8), P. 3490 - 3490

Published: April 17, 2023

Autonomous vehicles in highway driving scenarios are expected to become a reality the next few years. Decision-making and motion planning algorithms, which allow autonomous predict tackle unpredictable road traffic situations, play crucial role. Indeed, finding optimal decision all different is challenging task due large complex variability of scenarios. In this context, aim work design an effective hybrid two-layer path architecture that, by exploiting powerful tools offered emerging Deep Reinforcement Learning (DRL) combination with model-based approaches, lets properly behave conditions and, accordingly, determine lateral longitudinal control commands. Specifically, DRL-based high-level planner responsible for training vehicle choose tactical behaviors according surrounding environment, while low-level converts these choices into actions be imposed through optimization problem based on Nonlinear Model Predictive Control (NMPC) approach, thus enforcing continuous constraints. The effectiveness proposed hierarchical hence evaluated via integrated vehicular platform that combines MATLAB environment SUMO (Simulation Urban MObility) simulator. exhaustive simulation analysis, carried out non-trivial scenarios, confirms capability strategy

Language: Английский

Exploring Semantic Prompts in the Segment Anything Model for Domain Adaptation DOI Creative Commons
Ziquan Wang,

Yongsheng Zhang,

Zhenchao Zhang

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(5), P. 758 - 758

Published: Feb. 21, 2024

Robust segmentation in adverse weather conditions is crucial for autonomous driving. However, these scenes struggle with recognition and make annotations expensive, resulting poor performance. As a result, the Segment Anything Model (SAM) was recently proposed to finely segment spatial structure of provide powerful prior information, thus showing great promise resolving problems. SAM cannot be applied directly different geographic scales non-semantic outputs. To address issues, we propose SAM-EDA, which integrates into an unsupervised domain adaptation mean-teacher framework. In this method, use “teacher-assistant” model semantic pseudo-labels, will fill holes fine given by generate pseudo-labels close ground truth, then guide student learning. Here, helps distill knowledge. During testing, only used, greatly improving efficiency. We tested SAM-EDA on mainstream benchmarks obtained more-robust model.

Language: Английский

Citations

4
A Multi-agent Reinforcement Learning Based Control Method for Connected and Autonomous Vehicles in A Mixed Platoon DOI

Yaqi Xu,

Yan Shi, Xiaolu Tong

et al.

IEEE Transactions on Vehicular Technology, Journal Year: 2024, Volume and Issue: 73(11), P. 16160 - 16172

Published: June 18, 2024

Language: Английский

Citations

4
An integrated MCDM approach for enhancing efficiency in connected autonomous vehicles through augmented intelligence and IoT integration DOI Creative Commons
Saeid Jafarzadeh Ghoushchi, Sina Shaffiee Haghshenas,

Sahand Vahabzadeh

et al.

Results in Engineering, Journal Year: 2024, Volume and Issue: 23, P. 102626 - 102626

Published: July 27, 2024

The rapid advancement of Connected Autonomous Vehicles (CAVs) equipped with self-powered sensors is poised to revolutionize road safety, efficiency, and the quality travel. However, effective integration these technologies within dynamic environments poses significant challenges, highlighting need for innovative multi-criteria decision-making (MCDM) approaches optimize their deployment. This study tries solve problem by proposing an MCDM method that uses fuzzy sets evaluate rank different scenarios better performance augmented intelligence Internet Things (IoT) in CAVs. research two key techniques: Fuzzy Logarithm Incremental Weights (F-LMAW) criteria evaluation Fermatean Weighted Aggregated Sum Product Assessment (FF-WASPAS) scenario evaluation. To ensure reliability accuracy results, a sensitivity analysis was conducted. confirmed effectiveness proposed approach. study's results showed third (the IoT urban areas via sensors) got highest score, which shows how important it compared other choices. obtained highlight importance integrating enhance public transportation autonomous vehicles sensors.

Language: Английский

Citations

4
Simulation Study of Deep Belief Network-Based Rice Transplanter Navigation Deviation Pattern Identification and Adaptive Control DOI Creative Commons

Xianhao Duan,

Peng Fang,

Neng Xiong

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(2), P. 790 - 790

Published: Jan. 15, 2025

The navigation field of agricultural machinery has entered the intelligent stage, but control performance paddy represented by rice transplanters is not stable in complex environments. Therefore, this study proposes a method to identify deviation patterns based on Deep Belief Network (DBN) and designs an adaptive preview distance driver model for each pattern. Among them, pattern identification two-stage algorithm. First, determine whether current status abnormal. Then, classification was refined different abnormal states. divided into two levels. main regulator calculates dynamic according state variable; sub-regulator adjustment value degree. In test method, all models show excellent stability accuracy, speed algorithm meets high frequency transplanter system. algorithm, compared with static distance, proposed can effectively suppress disturbance navigation.

Language: Английский

Citations

0
Safeguarding connected autonomous vehicle communication: Protocols, intra- and inter-vehicular attacks and defenses DOI
Mohammed Aledhari, Rehma Razzak, Mohamed Rahouti

et al.

Computers & Security, Journal Year: 2025, Volume and Issue: unknown, P. 104352 - 104352

Published: Jan. 1, 2025

Language: Английский

Citations

0
An autonomous driving decision-making framework for joint prediction and planning in unsignalized intersection scenarios DOI
Shupei Zhang,

P P Sun,

Ying Pang

et al.

Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 19, 2025

Unsignalized intersections present one of the most challenging environments in autonomous driving due to their complex traffic scenarios. Safely and efficiently navigating these uncertain settings remains a significant research hurdle. To tackle this issue, paper proposes an End-to-End Autonomous Driving Decision Framework (EJPP) based on interactive fusion prediction planning modules. The framework accurately predicts future trajectories surrounding vehicles facilitate optimal path planning. EJPP consists module integrates vehicle acceleration as implicit behavioral intent utilizes Pearson correlation coefficients comprehensively consider complete information interactions among vehicles, thereby mitigating potential trajectory uncertainties. Moreover, temporal attention mechanism is incorporated capture critical features from historical enhance accuracy. Within module, are planned separately lateral longitudinal directions using Frenet coordinate system. By integrating dynamics into cost functions encompassing safety, comfort, efficiency, both soft hard constraints designed optimize route. proposed validated through closed-loop training across various flow scenarios assess its performance. Results indicate that enables traverse unsignalized safely efficiently.

Language: Английский

Citations

0
An efficient function placement approach in serverless edge computing DOI

Atiya Zahed,

Mostafa Ghobaei‐Arani, Leila Esmaeili

et al.

Computing, Journal Year: 2025, Volume and Issue: 107(3)

Published: Feb. 21, 2025

Language: Английский

Citations

0
Robust Control of Autonomous Robotic Vehicles Using Nonlinear Model Predictive Controller DOI Creative Commons
Aliasghar Arab

IntechOpen eBooks, Journal Year: 2025, Volume and Issue: unknown

Published: March 4, 2025

The proposed chapter will explore the application of robust control methods for electric connected autonomous robotic vehicles (EACVs) with a focus on techniques. As demand mobility increases, ensuring their safety, reliability, and efficiency becomes paramount. This delve into challenges solutions associated EACVs, particularly under varying operational conditions that cause uncertainties. By leveraging learning strategies, aims to demonstrate how these can enhance performance resilience EACVs. Key topics include integration nonlinear trajectory following, handling disturbances parameter variations, system robustness. also present case studies simulations illustrate effectiveness strategies in real-world scenarios. comprehensive overview provide valuable insights researchers, engineers, practitioners involved development deployment vehicle technologies.

Language: Английский

Citations

0
Data-driven Koopman model predictive control for the integrated thermal management of electric vehicles DOI
Youyi Chen, Kyoung Hyun Kwak, Dohoy Jung

et al.

Control Engineering Practice, Journal Year: 2025, Volume and Issue: 160, P. 106323 - 106323

Published: March 28, 2025

Language: Английский

Citations

0
An Optimal Scheduling Model for Connected Automated Vehicles at an Unsignalized Intersection DOI Creative Commons
Wei Bai,

Chengxin Fu,

Bin Zhao

et al.

Algorithms, Journal Year: 2025, Volume and Issue: 18(4), P. 194 - 194

Published: April 1, 2025

The application of connected automated vehicles (CAVs) provides new opportunities and challenges for optimizing controlling urban intersections. To avoid collisions in conflicting directions at intersections improve the efficiency intersections, an optimal scheduling model CAVs unsignalized intersection is proposed. develops a linear programming vehicle timing with minimum average delay within optimization time window as objective safe interval to pass through constraint. A rolling algorithm designed solution. Finally, effects different traffic demand conditions on results are investigated based numerical simulation experiments. show that both proposed Gurobi solver can significantly reduce compared first-come-first-served (FCFS) control method, by 76.22% most. Compared solver, solution ensure effect greatest extent. Therefore, provide theoretical support managing

Language: Английский

Citations

0