Particle swarm optimization (PSO) often stagnates in photovoltaic (PV) parameter extraction due to premature convergence and velocity decay. This paper proposes Random Reinitialized PSO (R²-PSO), which periodically reinitializes particle positions and velocities to restore diversity and sustain exploration without disrupting convergence. Performance is evaluated using the run-length distribution (RLD), a probabilistic benchmark that jointly characterizes convergence dynamics and steady-state success. In PV parameter extraction, R²-PSO achieves RMSE = 10⁻⁹, 100% success over 100 runs, and convergence in ≈1.05M function evaluations. Experimental validation on multi-crystalline (SM55), monocrystalline (S75), and thin-film (ST40) modules shows close agreement with manufacturer datasheets. Generalization capability is confirmed through the CEC-2022 single-objective benchmark suite, where R²-PSO attains the lowest total rank and 1st overall position among five competitive PSO-based algorithms, securing best-mean performance on 8 out of 12 functions. Statistical comparisons across both application and benchmark domains indicate uniformly lower error metrics (SSE, MAE, MSE, AE, SE) relative to competing methods. Sensitivity analyses further demonstrate robustness, maintaining 100% success under variations in the VTR threshold, population size, maximum velocity, and epoch initialization frequency. These results establish R²-PSO as an accurate, reliable, and computationally efficient optimizer for PV parameter extraction, while highlighting RLD as a rigorous standard for evaluating stochastic search algorithms.

The requirement of sustainable mobility and a clean environment has accelerated the development and adoption of electric vehicles (EVs) and hybrid electric vehicles (HEVs) as an alternative, practical and promising solution against conventional vehicles globally. Such alternative energy vehicles not only provide a critical solution to mitigate fossil fuel dependency and reduce greenhouse gas emissions, but also contribute to producing an energy-efficient transportation system. However, the operational performance, efficiency, and cost-effectiveness of EVs and HEVs are hugely dependent on their powertrain architectures, selection of traction motors and associated control techniques. This paper systematically compares major hybrid architectures: series, parallel, and series–parallel, plug-in, as well as battery and fuel cell electric vehicle platforms, highlighting trade-offs in component sizing, cost, and system integration complexity. The paper critically analyses traction motor technologies with respect to torque–speed characteristics, efficiency behavior, material constraints, and power density. A detailed comparative assessment of traction motor technologies is presented. Furthermore, classical and advanced motor control strategies, including field-oriented control (FOC), direct torque control (DTC), model predictive control (MPC) and AI-enhanced control frameworks, are evaluated with respect to transient performance, robustness, computational requirements, and scalability. The review identifies key technological milestones, emerging next-generation drive technologies, existing limitations, and unresolved research challenges. Finally, critical research gaps and future development pathways are articulated to support the advancement of high-efficiency, reliable, and cost-effective EV/HEV powertrain systems.

This paper presents an adaptive population size (NP)–based accelerated Particle Swarm Optimization (AAPSO) algorithm for duty cycle–based maximum power point tracking (MPPT) in photovoltaic (PV) systems. The proposed method directly modulates the duty cycle of a DC–DC converter, enabling rapid and precise adjustments to the maximum power point (MPP) under both uniform and partial shading conditions. AAPSO enhances conventional PSO by adopting a social‐only variant and an adaptive Population size (NP) mechanism that begins with a large population for exploration and gradually reduces it to balance exploration and exploitation. To ensure robustness, the algorithm is executed 100 times, and performance is analyzed using statistical metrics and run‐length distribution (RLD). Simulation results demonstrate approximately 99.8% tracking efficiency with a 100% tracking accuracy across all runs, while convergence counts are reduced nearly threefold compared to conventional Particle Swarm Optimization (CPSO) and two recent adaptive PSO‐based MPPT methods from the literature. Experimental validation using a Ćuk converter prototype further confirms its practical feasibility. Overall, this study contributes an adaptive, duty cycle–based constrained PSO framework that integrates robustness, scalability, and statistical reliability for MPPT in large‐scale PV systems.

This paper proposes a new fuzzy logic-based rotor flux model reference adaptive system (FLC-MRAS) for rotor speed estimation in induction motor drives, replacing the constant-gain PI controller used in conventional MRAS schemes. The proposed observer simultaneously incorporates both rotor flux and electromagnetic torque errors to enhance estimation accuracy and robustness against load torque disturbances. A nonlinear Mamdani-type fuzzy logic controller (FLC) with two inputs and one output, employing triangular membership functions and seven fuzzy sets, is adopted. The effectiveness and useful operational performance of the proposed approach is examined through extensive simulation cases under various vehicle speed driving profiles and load torque conditions using an indirect vector-controlled induction motor drive. In order to investigate the effective operational performance of a speed estimator, different cases are prepared according to the vehicle requirements. To examine the robustness of the proposed observer under realistic operating conditions, rotor resistance variation is incorporated into the simulation framework. This approach allows assessment of MRAS performance under practical nonlinearities and parameter uncertainties encountered in real applications. Comparative results demonstrate superior speed regulation and speed tracking, reduced estimation error, and faster convergence of the adaptive tuning signal for better speed estimation compared to the PI-MRAS observer. The proposed scheme provides the suitable choice of traction drive adoption for electric vehicle (EV) applications.

This study proposes a novel modeling approach based on an Adaptive Lightweight Trust-Region Reflective (AL-TRF) method for estimating all nine three-diode model (TDM) parameters under arbitrary irradiance and temperature conditions using only manufacturer datasheet information. The proposed AL-TRF scheme simultaneously identifies all nine model parameters under varying operating conditions. Its performance is validated using three PV modules representing multi-crystalline (SM55), mono-crystalline (ST40), and thin-film (S75) technologies. The results demonstrate accurate reproduction of prescribed datasheet characteristics and measured current–voltage (I–V) characteristics over a wide range of irradiance and temperature conditions. Compared with STC-only parameter transport, condition-wise identification reduces the pooled RMSE by 4.02%, 8.43%, and 26.68% for ST40, SM55, and S75, respectively. For the SM55 and S75 modules, AL-TRF achieves lower mean-condition RMSE than Harris Hawks Optimization (HHO) and Sunflower Optimization (SFO), with reductions of 1.75% and 18.76% relative to HHO and 66.24% and 24.13% relative to SFO, respectively. For ST40, despite slightly lower prediction errors obtained by the reference methods, AL-TRF retains a distinct modeling advantage through complete condition-wise identification of all nine TDM parameters. This comprehensive capability is achieved with substantially lower computational effort. In the single-STC SM55 benchmark, AL-TRF reduces execution time by 79.31% and 83.90% relative to HHO and SFO, respectively. For the complete modeling protocols, the corresponding reductions are 67.15% and 79.32% for ST40 and 82.12% and 82.12% for S75. Overall, AL-TRF provides a superior balance of predictive accuracy, complete condition-wise nine-parameter identification, and computational efficiency, establishing it as an effective datasheet-driven approach for reliable TDM reconstruction under widely varying operating conditions.

This article presents a computationally efficient maximum power point tracking (MPPT) strategy for high‐power permanent‐magnet synchronous generator (PMSG)‐based wind energy conversion systems (WECSs). A reduced‐input fuzzy controller (RFC) is developed by transforming a conventional dual‐input fuzzy controller (DFC) into a single‐input structure through a signed‐distance formulation that combines the error and change‐of‐error variables into a unified control signal. The proposed transformation reduces the fuzzy rule base from 49 to 7 rules, corresponding to an 85.7% reduction in inference requirements, while preserving the control characteristics of the original DFC. The controller is implemented on a 2 MW WECS employing a three‐phase interleaved boost converter (IBC) and evaluated under identical operating conditions against the DFC and previously benchmarked heuristic MPPT methods. Comprehensive simulation studies are conducted under varying wind profiles, including turbulent and low‐wind conditions, ±5% sensor noise, ±10% converter parameter variations, reference changes, and transitions between continuous and discontinuous conduction modes (DCMs). Performance is assessed using power‐tracking characteristics, ripple analysis, integral error indices, and computational complexity measures. Results show that the RFC maintains MPPT performance comparable to the conventional DFC while reducing controller execution time from 250 to 50 ms, corresponding to an 80% reduction in computational time. In addition, the RFC achieves a 14.3% reduction in power ripple and maintains robust operation under measurement noise, parameter uncertainty, and varying operating conditions. These findings confirm that the proposed RFC provides a practical and resource‐efficient MPPT solution suitable for real‐time embedded implementation in large‐scale wind energy systems.