Resilient Reinforcement Learning for Voltage Control in an Islanded DC Microgrid Integrating Data-Driven Piezoelectric

Machines, Год журнала: 2024, Номер 12(10), С. 694 - 694

Опубликована: Окт. 1, 2024

This research study presents a resilient control scheme for an islanded DC microgrid (DC MG) integrating solar photovoltaic (PV), battery storage (BESS), and piezoelectric (PE) energy harvesting modules. The (MG) case represents hub designed to provide electricity lighting systems in transportation, roads, other infrastructure. To enhance practicality, the PE is modeled using real data captured from traffic simulator. proposed reinforcement learning (RL) method was tested against four severe unexpected failure scenarios, including short circuit at load side, sudden change of load, open circuit, converter failure. performance controller quantitatively compared with conventional PI controller. results show marginal improvement one scenario significant three, suggesting that robust candidate microgrids high levels uncertainty, such as those involving harvesters.

Язык: Английский

---

ADP-based fault-tolerant consensus control for multiagent systems with irregular state constraints

Neural Networks, Год журнала: 2024, Номер 180, С. 106737 - 106737

Опубликована: Сен. 14, 2024

Язык: Английский

---

Event‐Based Prescribed‐Time Containment Control for Multiple Euler–Lagrange Systems via Super‐Twisting Sliding Mode

International Journal of Robust and Nonlinear Control, Год журнала: 2025, Номер unknown

Опубликована: Янв. 12, 2025

ABSTRACT This article designs a distributed dynamic event‐triggered prescribed‐time sliding mode controller to tackle the containment control problem of multiple Euler–Lagrange systems (MELSs) with external disturbances and strict requirements on response speed. Unlike previous super‐twisting methods that achieve finite‐time control, this paper replaces constant term in time scaling function variable incorporates it into design mechanism (DETM), achieving (PTC), enhancing system robustness reducing updates. Based structure theory PTC theory, non‐linear manifold is devised for purpose ensuring convergence error zero within prescribed time. The (PTSTC) protocol designed ensure reachability manifold, while attenuating chattering during process. can be set arbitrarily. To minimize frequency updates losses actuator, DETM deployed controller‐to‐actuator channel. Moreover, absence Zeno phenomenon derived. Error signals are proven converge using Lyapunov stability criteria. Ultimately, simulation manipulator indicates effectively drives zero. Meanwhile, number triggers reduced by more than 50% when framework involves compared static mechanisms.

Язык: Английский

---

Robust Adaptive Control for High‐Order Nonlinear Systems With Unknown Upper Bound Uncertainties Based on Fully Actuated System Approaches and Multi‐Objective Optimization

International Journal of Robust and Nonlinear Control, Год журнала: 2025, Номер unknown

Опубликована: Янв. 21, 2025

ABSTRACT This study focuses on the design of a robust adaptive (RA) controller and tracking (RAT) for high‐order nonlinear systems that have unknown upper bound time‐varying uncertainties. Utilizing fully actuated (HOFA) system approaches Lyapunov stability theory, proposed RA RAT controllers are capable overcoming uncertainties, while also automatically adjusting parameter estimates associated with The ensure both state closed‐loop estimation error vector converge globally to bounded ellipsoid. Furthermore, degrees freedom provided by HOFA method is utilized regional pole assignment, an objective function designed enhance performance. Finally, effectiveness practicality control demonstrated through simulation examples, including two‐link manipulator.

Язык: Английский

---

Date-driven deadbeat sliding mode control using input/output differences

Journal of the Franklin Institute, Год журнала: 2025, Номер unknown, С. 107622 - 107622

Опубликована: Март 1, 2025

Язык: Английский

---

Machine learning in parameter estimation of nonlinear systems

The European Physical Journal B, Год журнала: 2025, Номер 98(4)

Опубликована: Апрель 1, 2025

Abstract Accurate parameter estimation in nonlinear dynamical systems remains a fundamental challenge due to noise, limited data, and model complexity. Traditional methods, such as gradient-based optimization least squares (NLS), often struggle under real-world multiplicative exhibiting sensitivity outliers high computational demands. This study introduces neural network framework integrating the Huber loss function achieve robust efficient estimation. Applied canonical systems, including damped oscillators, van der Pol Lotka–Volterra models, chaotic Lorenz dynamics, proposed method demonstrates superior accuracy resilience noise. Notably, it maintains sub- $$1.2\%$$ 1.2 % relative errors for key parameters system, significantly outperforming NLS, which diverges with exceeding $$12\%$$ 12 identical noise conditions. The use of SiLU activation improves convergence, yielding statistically significant reductions ( $$p < 0.01$$ p < 0.01 ). Furthermore, operates up $$8\times $$ 8 × faster than conventional techniques while reducing root-mean-square error by over $$99.9\%$$ 99.9 high-noise regimes. These results establish robust, data-driven approach complex bridging machine learning physics enabling real-time applications noisy environments. Graphic

Язык: Английский

---

Event-driven intelligent fault-tolerant containment control for nonlinear multiagent systems with unknown disturbances

International Journal of Systems Science, Год журнала: 2024, Номер unknown, С. 1 - 16

Опубликована: Окт. 4, 2024

Язык: Английский

---

Data-Driven Prescribed Performance Platooning Control Under Aperiodic Denial-of- Service Attacks

Mathematics, Год журнала: 2024, Номер 12(21), С. 3313 - 3313

Опубликована: Окт. 22, 2024

This article studies a data-driven prescribed performance platooning control method for nonlinear connected automated vehicle systems (CAVs) under aperiodic denial-of-service (DoS) attacks. Firstly, the dynamic linearization technique is employed to transform CAV system into an equivalent linearized data model. Secondly, improve system’s transient performance, transformation (PPT) scheme proposed constrained output unconstrained one. In addition, attack compensation mechanism designed reduce adverse impact. Combining PPT and mechanism, adaptive achieve vehicular tracking task. Lastly, merits of developed are illustrated by actual simulation.

Язык: Английский

---

Distributed‐filter‐based double event‐triggered formation control for stochastic nonlinear MASs under switching topologies

International Journal of Robust and Nonlinear Control, Год журнала: 2024, Номер unknown

Опубликована: Сен. 11, 2024

Abstract This article investigates a formation control method for stochastic nonlinear multi‐agent systems (MASs) under switching topologies. To reduce the communication bandwidth occupancy, two event‐triggered mechanisms of sensor‐to‐controller and controller‐to‐actuator network channels are proposed. Taking advantage neural networks approximation capability, dynamic high‐gain observer is introduced to estimate unmeasured states tackle non‐differentiable issue triggered output signal. Furthermore, it should be noted that distributed filter employed handle discontinuous local reference signal resulting from By using topology information, generates differentiable design virtual controller. Concomitantly, first‐order implemented avoid problem “explosion complexity.” Through stability analysis, proven designed controller achieves boundedness in probability all signals MASs. Ultimately, simulation performed confirm viability approach.

Язык: Английский

---

Consensus-Based Power System State Estimation Algorithm Under Collaborative Attack

Sensors, Год журнала: 2024, Номер 24(21), С. 6886 - 6886

Опубликована: Окт. 27, 2024

Due to its vulnerability a variety of cyber attacks, research on security for power systems has become especially crucial. In order maintain the safe and stable operation systems, it is worthwhile gain insight into complex characteristics behaviors attacks from attacker’s perspective. The consensus-based distributed state estimation problem investigated subject collaborative attacks. describe such attack behaviors, denial service (DoS) model hybrid remote terminal unit (RTU) phasor measurement (PMU) measurements, false data injection (FDI) neighboring information, are constructed. By integrating these two types models, different estimator designed accurately estimate system under Then, through Lyapunov stability analysis theory, sufficient condition provided ensure that proposed stable, suitable consensus matrix devised. Finally, confirm viability efficacy suggested algorithm, simulation experiment an IEEE benchmark 14-bus carried out.

Язык: Английский

---