ABSTRACTThis article investigates the fixed‐time optimal tracking control problem for the single‐link manipulator time‐delay system with unknown disturbance. An improved fixed‐time extended state observer is established, which not only estimates the disturbance existing in the system but also estimates the state information of the system. Subsequently, the reinforcement learning (RL) approach with a critic‐actor framework is utilized to construct the optimal fixed‐time controller. During the RL process, a modification term is introduced to prevent the occurrence of the training termination, in which critic and actor algorithms can converge within a fixed‐time interval. Finally, an appropriate Lyapunov‐Krasovskii function is designed to compensate for the state delay. With the proposed optimal control scheme, all signals achieve the fixed‐time convergence, which ensures that the system output effectively tracks the desired trajectory. The effectiveness of the proposed strategy is verified through the simulation example.

ABSTRACTThis article proposes a parallel algorithm of stochastic model predictive control (SMPC) for linear discrete‐time systems. The proposed algorithm uses multiple threads to solve respective optimization problems at different time steps in parallel, without inter‐thread communication during each optimization. Each thread sequentially solves a stochastic optimal control problem (SOCP) for optimizing closed‐loop performance against future disturbances and a partial minimization problem for optimizing a subset of decision variables of the SOCP. The partial minimization problem computes the control input according to the last measured state to satisfy chance constraints and can be solved much faster than the original SOCP. Consequently, the proposed algorithm can reduce the implementable minimum sampling period compared to the original SMPC. A numerical example of controlling room temperature showed that the proposed algorithm achieves control performance almost identical to that of closed‐loop SMPC, which optimizes closed‐loop performance. However, the implementable minimum sampling period of the proposed algorithm is much shorter than that of the closed‐loop SMPC, even with parallelization overhead.

ABSTRACTThe rapid combination of the electric vehicles into the recent transportation prefers very efficient charging systems involved in grid conditions. This increasing adoption of electric vehicles demands a dynamic and intelligent framework to control charging, confirming optimal grid performance, load balancing, and cost efficiency. In this work, we developed an optimized deep learning framework using the combined structure of Whale‐Optimized Neuro‐Fuzzy Classification for controlling electric vehicle charging within the grid. The increasing involvement of electric vehicle's produces new difficulties, involving overload of grid, and energy management, in real‐time and optimized decision making process. The Whale‐Optimized Neuro‐Fuzzy Classification method uses the hybrid abilities of neuro‐fuzzy systems, integrates the capability of the neural networks and also optimize using a Whale Optimization Algorithm for improved accuracy and efficiency. This proposed method maintains the charging and discharging process of electric vehicles, which have several factors like grid load, priorities of the vehicle and preferences of the user. The neuro‐fuzzy system combines deep convolutional neural network and fuzzy logic to predict charging patterns, considering user preferences, grid demand, renewable energy availability, and so on. This method contribute to the enhancement of energy systems, denotes the future requirement of future smart cities. Evaluation metrics included power usage, energy loss, cost, and processing speed. Analysis of experimental results revealed the accuracy of 99% for the proposed method compared to existing techniques.

ABSTRACTTo address the challenge of parallel parking in narrow environments, a multi‐segment path planning method is proposed for autonomous unmanned vehicles (AUVs), which integrates circular arc and dual clothoid curves, ensuring continuous path curvature and minimal longitudinal space occupancy. First, considering the narrow environmental constraints and the parameters of AUV, the planning for the preparatory phase path and the parking‐in phase path are completed, respectively. The role of the preparatory phase path is to guide the AUV to the target point, and the parking‐in phase path can guide the AUV into the target parking space. Then, by using the particle swarm optimization (PSO) algorithm with a penalty function framework, a multi‐constraint nonlinear programming path function model is established to seek out the parking path. Subsequently, the pure pursuit (PP) algorithm is employed to control the movement of AUV along the planned path. Finally, the simulation result and real‐vehicle experiment are conducted, which show that the parking path planning method proposed in this article enables AUV to safely complete parking operations in the narrow environment.

ABSTRACTThis paper develops an adaptive neural optimized control strategy for a class of switched strict‐feedback systems with unknown control gains and prescribed performances utilizing an adaptive projection method. To design the desired adaptive neural controller, we firstly utilize the actor‐critic architecture to formulate both optimized virtual and actual controllers within the backstepping framework. Secondly, this paper introduces an adaptive projection to eliminate the assumption that the unknown control coefficient cannot pass through zero. Moreover, control issues related to predefined time constraints and precision requirements can be addressed through error transformation. The proposed control strategy ensures that the tracking error converges to the predefined range within the allocated time. Additionally, all the signals of the controlled system remain bound. Finally, a set of comparative simulation experiments is discussed for three different predefined times, and the experimental results confirm the feasibility of the proposed control strategy.

ABSTRACTLithium‐ion Battery State of Health (SOH) estimation is essential for reliability and safety. Rechargeable lithium‐ion availability is decreased by thickening the solid electrolyte interphase (SEI) layer due to interactions between electrodes and electrolytes during cycles of discharging and charging. This paper proposes a hybrid approach, called LOA‐FENN, to reduce estimation errors by fusing the fully Elman neural network (FENN) with the lyrebird optimization algorithm (LOA). The LOA optimizes the neural network's weight during training, while the FENN predicts the SOH. Implemented in MATLAB, the LOA‐FENN approach achieved an estimation error of 1.08%, outperforming existing methods such as the hybrid genetic algorithm (HGA), box‐cox transformation (BCT), and particle filter algorithm (PFA), which showed errors of 1.38%, 1.28%, and 1.18%, respectively.

ABSTRACTBand dryers commonly used in industries consist of forced draft fans and heating elements. The heating coils used in these dryers are susceptible to voltage fluctuations, which might lead to poor performance of heating elements and high‐power consumption for operation. This manuscript proposes a novel approach for a setpoint tracking control scheme for the digital twin of the industrial finned heater. The proposed technique is a Butterfly Optimization Algorithm (BOA) for optimizing the Setpoint Tracking Proportional Integral Derivative (STPID) controller. The primary goal of the proposed method is to minimize error and provide a better alternative to find their system transfer function. The BOA technique is used to maximize the parameters of the STPID controller. On the MATLAB platform, the proposed technique is assessed and contrasted with other existing techniques. The simulation results show a decent control of temperature in the system using the OSTPID controller. The proposed method shows better results in all existingcontroller like Fractional Order Proportional Integral Derivative (FOPID) and STPID controllers. The output of OSTPID controller shows notable improvement with respect to settling time as 25.99% for a step input and 11.765% for random input, overshoot as 91.99% for a step input, and 92.53% for random input, RMS as 3.62% for a step input and 0.947% for random input and other time integral performance criteria like ISE, IAE, ITAE, etc. Based on the findings, it can be concluded that the proposed technique's based error is smaller than that of the other approaches.

ABSTRACTIn this article, we investigates the event‐triggered dissipative control problem of discrete‐time two‐dimensional (2‐D) switched systems with mixed delays. First, according to the evolution characteristics of the Fornasini–Marchesini model, a bidirectional event‐triggered mechanism (ETM) is adopted to reduce unnecessary calculation and control signal transmission. Second, an improved Lyapunov–Krasovskii function is selected, combined with Jensen inequalities and Abel lemma‐based finite‐sum inequalities, and according to the extended average dwell time method, sufficient conditions are given in the form of linear matrix inequalities to ensure the dissipative stability of the closed‐loop system under bidirectional ETM. Then, the corresponding state feedback controller and dynamic output feedback controller based on bidirectional ETM were designed to realize the dissipative control of the 2‐D switched systems. Finally, numerical examples verify the effectiveness of the controller design method.

ABSTRACTThe Reptile Search Algorithm (RSA) is a powerful modern optimization technique that effectively solves intricate problems across various fields. Despite its notable success, the local search aspect of RSA requires enhancement to overcome issues such as limited solution variety, a pattern of falling into local optimal traps, and the possibility of early convergence. In response to these challenges, this research introduces an innovative paradigm that melds the robust and time‐honoured local search technique, Simulated Annealing (SA), with RSA, christened henceforth as SARSA. This amalgamation aims to tackle the qualities of both strategies, synergistically improving their optimization capabilities. We utilize a broad and thorough assessment system to survey the viability and strength of SARSA. A comprehensive cluster of benchmark issues sourced from the CEC 2019 benchmark suite and an assorted set of real‐world challenges drawn from the CEC 2011 store is utilized as the test bed. This fastidiously curated testbed guarantees an intensive examination of SARSA's execution over a wide range of issues and complexities. Our observational discoveries substantiate that SARSA beats the foundational RSA and a few related calculations reported within the existing body of writing, in this manner setting up SARSA as a critical progression in optimization calculations. The prevalent execution illustrated by SARSA highlights its potential for broad application and underscores its utility in handling complex optimization issues viably.

ABSTRACTA multiple augmented Lyapunov function (MALF) is a function that includes more than one augmented matrix, allowing for a more precise evaluation of the relationships between different states within a system and facilitating the analysis of system stability. However, employing multiple augmented matrices can complicate the process of designing controllers. For this issue, this article proposes a controller synthesis algorithm to address the complexity of controller design resulting from multiple augmentation matrices. First, a memory state‐feedback control criterion with tuning matrices is derived by constructing an MALF. Second, a tuning matrix search algorithm is proposed to provide appropriate tuning matrices for the memory state‐feedback control criterion, leading to the determination of optimal controller gains. Finally, the effectiveness and advantage are verified by simulation of load frequency control for in the power system and control for a linear singular time‐delay system.