检索结果-内蒙古大学图书馆

Certificated Actor-Critic: Hierarchical Reinforcement Learning with control Barrier Functions for Safe Navigation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Xie, Junjun Zhao, Shuhao Hu, Liang Gao, Huijun Department of Automation School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen China Research Institute of Intelligent Control and Systems Harbin Institute of Technology Harbin China

control Barrier Functions (CBFs) have emerged as a prominent approach to designing safe navigation systems of robots. Despite their popularity, current CBF-based methods exhibit some limitations: optimization-based safe control techniques tend to be either myopic or computationally intensive, and they rely on simplified system models;conversely, the learning-based methods suffer from the lack of quantitative indication in terms of navigation performance and safety. In this paper, we present a new model-free reinforcement learning algorithm called Certificated Actor-Critic (CAC), which introduces a hierarchical reinforcement learning framework and well-defined reward functions derived from CBFs. We carry out theoretical analysis and proof of our algorithm, and propose several improvements in algorithm implementation. Our analysis is validated by two simulation experiments, showing the effectiveness of our proposed CAC algorithm. Copyright © 2025, The Authors. All rights reserved.

关键词： Adversarial machine learning

Stacked-Residual PINN for State Reconstruction of Hyperbolic systems

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Eshkofti, Katayoun Barreau, Matthieu Division of Decision and Control Systems Digital Futures KTH Royal Institute of Technology StockholmSE-100 44 Sweden

In a more connected world, modeling multi-agent systems with hyperbolic partial differential equations (PDEs) offers a potential solution to the curse of dimensionality. However, classical control tools need adaptation for these complex systems. Physics-informed neural networks (PINNs) provide a powerful framework to fix this issue by inferring solutions to PDEs by embedding governing equations into the neural network. A major limitation of original PINNs is their inability to capture steep gradients and discontinuities in hyperbolic PDEs. This paper proposes a stacked residual PINN method enhanced with a vanishing viscosity mechanism. Initially, a basic PINN with a small viscosity coefficient provides a stable, low-fidelity solution. Residual correction blocks with learnable scaling parameters then iteratively refine this solution, progressively decreasing the viscosity coefficient to transition from parabolic to hyperbolic PDEs. Applying this method to traffic state reconstruction improved results by an order of magnitude in relative L2 error, demonstrating its potential to accurately estimate solutions where original PINNs struggle with instability and low fidelity. Copyright © 2025, The Authors. All rights reserved.

关键词： Neural networks

Path Planning for Autonomous Underwater Vehicles in Complex Environments Based on a PRM* Algorithm

学校读者我要写书评

暂无评论

Path Planning for Autonomous Underwater Vehicles in Complex ...

IEEE SSD International Multi-Conference on systems, Signals and Devices

作者： Ali Arifi Raja Jarray Soufiene Bouallègue Research Laboratory in Automatic Control (LARA) of National Engineering School of Tunis National Engineering School of Gabès University of Gabès Gabès Tunisia Research Laboratory in Automatic Control (LARA) of National Engineering School of Tunis High Institute of Industrial Systems of Gabès Gabès Tunisia

ISBN: (数字)9798331542726

ISBN: (纸本)9798331542733

This paper proposes a Probabilistic Roadmap (PRM) planner for Autonomous Underwater Vehicles (AUVs) navigating in complex 3D environments. An improved version of PRM algorithms, called PRM*, is introduced to improve the capabilities of nodes specification, edges construction, and graph creation. For implementation, a pseudo-code for the proposed PRM* planner is provided. Numerical simulations are performed in different scenarios with increasing number of static obstacles and complexity. Evaluations of the generated collision-free paths are carried out using the Straight-Line Rate (SLR) and Computational Time (CT) metrics. Analyses and comparisons are performed to validate the significance and superiority of the proposed PRM* planner. Demonstrative results show that the PRM* algorithm outperforms the ordinary PRM one in terms of path compactness, smoothness, and collision avoidance.

关键词： Autonomous underwater vehicles Three-dimensional displays Navigation Probabilistic logic Numerical simulation Path planning Complexity theory Collision avoidance Vehicle dynamics Robots

Sparse Estimation of Inverse Covariance and Partial Correlation Matrices via Joint Partial Regression

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Erickson, Samuel Rydén, Tobias Division of Decision and Control Systems EECS KTH Royal Institute of Technology Sweden Lynx Asset Management AB Sweden

We present a new method for estimating high-dimensional sparse partial correlation and inverse covariance matrices, which exploits the connection between the inverse covariance matrix and linear regression. The method is a two-stage estimation method wherein each individual feature is regressed on all other features while positive semi-definiteness is enforced simultaneously. We provide statistical rates of convergence for the proposed method which match, and improve upon, the state-of-the-art for inverse covariance and partial correlation matrix estimation, respectively. We also propose an efficient proximal splitting algorithm for numerically computing the estimate. The effectiveness of the proposed method is demonstrated on both synthetic and real-world data. © 2025, CC BY.

关键词： Covariance matrix

Closed-Loop Neural Operator-Based Observer of Traffic Density

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Harting, Alice Johansson, Karl Henrik Barreau, Matthieu Division of Decision and Control Systems Digital Futures KTH Royal Institute of Technology StockholmSE-100 44 Sweden

We consider the problem of traffic density estimation with sparse measurements from stationary roadside sensors. Our approach uses Fourier neural operators to learn macroscopic traffic flow dynamics from high-fidelity microscopic-level simulations. During inference, the operator functions as an open-loop predictor of traffic evolution. To close the loop, we couple the open-loop operator with a correction operator that combines the predicted density with sparse measurements from the sensors. Simulations with the SUMO software indicate that, compared to open-loop observers, the proposed closed-loop observer exhibit classical closed-loop properties such as robustness to noise and ultimate boundedness of the error. This shows the advantages of combining learned physics with real-time corrections, and opens avenues for accurate, efficient, and interpretable data-driven observers. Copyright © 2025, The Authors. All rights reserved.

关键词： Traffic control

Optimal Design and Tuning of TID Power System Stabilizer for Multi-machine Power System Using Dandelion Optimizer

学校读者我要写书评

暂无评论

Optimal Design and Tuning of TID Power System Stabilizer for...

International Conference on Artificial Intelligence in Renewable Energetic systems, IC-AIRES 2024

作者： Hattabi, Intissar Kheldoun, Aissa Bradai, Rafik Khettab, Soufian SET Laboratory Blida 1 University Faculty of Technology Electrical and Control Department Blida Algeria Laboratory of Signals and Systems Institute of Electrical and Electronic Engineering University M’hamed Bougara Boumerdes35000 Algeria LATSI Blida 1 University Faculty of Technology Electronics Department Blida09000 Algeria

ISBN: (纸本)9783031803000

In this study, we explore the enhancement of stability in a four-machine power system through the optimized tuning of Tilt-Integral-Derivative Power System Stabilizer (TID-PSS) parameters. We introduce a novel application of the Dandelion Optimizer (DO) algorithm, designed to provide robust damping and stability under fault conditions. To improve the effectiveness of the DO algorithm, we compare its performance with several established optimization techniques, including Particle Swarm Optimization (PSO), HGS (Hunger Games Search), HBA (Honey Badger Algorithm), EO (Equilibrium Optimizer), and AVOA (African Vulture Optimization algorithm. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Particle swarm optimization (PSO)

The influence mechanism of transmission stiffness on mechanical resonance of aircraft rudder servo system

学校读者我要写书评

暂无评论

The influence mechanism of transmission stiffness on mechani...

2024 International Conference on Aerospace, Mechanical and Materials Engineering, AMME 2024

作者： Xu, Zhiwei Tang, Dejia Xu, Dufeng Fang, Li Fu, Yiyong School of Mechanical Engineering Nanjing University of Science and Technology Nanjing210094 China School of Aeronautics and Astronautics Shanghai Jiao Tong University Shanghai200240 China Shanghai Aerospace Control Technology Institute Shanghai201109 China Shanghai Engineering Research Center of Servo Systems Shanghai201109 China

Regarding the mechanical resonance problem caused by the coupling of flexible characteristics and transmission stiffness, this typical problem is faced by aircraft rudder servo systems under the requirement of lightweight design. A model of the aircraft rudder servo system considering flexible coupling links is established, and the influence of transmission stiffness on the open-loop transfer function of the rudder servo system is analyzed. Increasing the transmission stiffness increases both the anti-resonance frequency and the resonance frequency of the rudder servo system, and the system performance changes depending on the change in open-loop gain margin. When the load-motor rotational inertia and product are small, and the resonant frequency is still much lower than the natural frequency of the current loop after increasing the transmission stiffness, the system open-loop gain margin will decrease and mechanical resonance will intensify. When the load-motor rotational inertia and product are relatively large, increasing the transmission stiffness to make the resonant frequency close to or even greater than the natural frequency of the current loop will increase the system's open-loop gain margin and weaken mechanical resonance. The correctness of the research conclusion was verified through system simulation and experimental validation. © 2025 institute of Physics Publishing. All rights reserved.

关键词： Natural frequencies

Development of automatic Cartesian Robot Cannabis Seeder using LabVIEW

学校读者我要写书评

暂无评论

Development of Automatic Cartesian Robot Cannabis Seeder usi...

International ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI-NCON)

作者： Rujipan Kosarat Keattipoom Sangwai Anan Wongjan Department of Software Engineering Faculty of Engineering Rajamangala University of Technology Lanna Chiangmai Thailand Department of Electronic Engineering and Automatic Control Systems Faculty of Engineering Rajamangala University of Technology Lanna Chiangmai Thailand

ISBN: (数字)9798331543273

ISBN: (纸本)9798331543280

This article describes the invention of autonomous cannabis seeding equipment to reduce contamination and planting time. The automated cannabis seeder employs an NI myRIO control board to operate a y-axis stepper motor attached to a ball screw to move along the breadth of each row of 14 seed trays. Soil drilling and burial Drill all holes at the same depth. This promotes seed germination and lowers human-seed germ contamination. These 7 soil drill heads and 7 seeders use 3D printers to design and shape the workpiece. An 800-watt air pump smokes the cannabis seeds. Linear electric motors function in two rows in the z-axis. The buffer reverses the drowning process from row 14 to the first row, starting from the 3rd row and continuing through the final series of steps until 14 rows in the y-axis are complete. The hardware controls all program activities in LabVIEW. We tested the automated cannabis seeding machine 100 times on average, sowing cannabis seeds into 98 seed trays and dropping 93 seeds at 94.80% efficiency in 10.38 minutes. A 16 percent faster machine than human work is available.

关键词： Technological innovation Motor drives Shape Soil Pulse width modulation Three-dimensional printing Seeds (agriculture) Printers Robots Contamination

Demonstration of the Dyna Reinforcement Learning Framework for Reactive Close Proximity Operations

学校读者我要写书评

暂无评论

Demonstration of the Dyna Reinforcement Learning Framework f...

AIAA Science and technology Forum and Exposition, AIAA SciTech Forum 2025

作者： Majumdar, Ritwik Albee, Keenan Sternberg, David C. Jia-Richards, Oliver Department of Aerospace Engineering University of Michigan Ann ArborMI United States Payload/Instrument Pointing Control Systems Engineering Group Jet Propulsion Laboratory California Institute of Technology PasadenaCA United States Maritime and Multi-Agent Autonomy Group Jet Propulsion Laboratory California Institute of Technology PasadenaCA United States

ISBN: (数字)9781624107238

ISBN: (纸本)9781624107238

Lessons from the International Space Station (ISS) emphasize the necessity of exterior inspection for anomaly detection and maintenance, but current methods rely on costly and limited human extravehicular activities and robotic arms. Deployable free-flying small spacecraft offer a flexible, autonomous solution, capable of comprehensive exterior inspections without human involvement. However, the safety of these spacecraft during close proximity operations remains a concern, particularly given uncertain variability in thruster performance. This paper presents SmallSat Steward, a reactive and integrated architecture for online model learning and trajectory planning based on the Dyna reinforcement learning architecture. By combining model-based planning and direct reinforcement learning, Dyna offers a potentially flexible and computationally-efficient solution capable of adapting to changes in thruster performance and other system uncertainties. Preliminary results in both simulation and hardware environments demonstrate the potential of this architecture to successfully regulate position under single and double thruster failures. In simulation, the Dyna-based controller outperformed an PD-LQR controller in ∼70% of all cases. On hardware, Dyna was able to eliminate the steady state error caused by thruster failures. © 2025, American institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved.

关键词： Space stations