To address the challenges of difficult quantitative design and potential coil mismatch in auxiliary coils within wireless power transfer systems, a data-driven parameter optimization method based on multi-objective particle swarm optimization (MOPSO) was proposed. First, based on the inductor–capacitor–capacitor series (LCC-S) compensation topology, a mechanism-based analysis was conducted, establishing coil side length A and number of turns N as core optimization variables. Subsequently, a collaborative optimization framework integrating “parametric simulation–surrogate modeling–active learning” was established. An offline fingerprint database was constructed via finite element simulation, and a high-accuracy surrogate model was developed using a kernel ridge regression ensemble approach. Active learning strategies were employed to adaptively augment data points and mitigate uncertainty. Finally, the multi-objective particle swarm optimization (MOPSO) algorithm was applied to identify the Pareto-optimal solution set. Experimental results reveal that the optimized auxiliary coil parameters achieved positioning errors below 8 mm at all test points. The maximum positioning error was significantly reduced by approximately 80% compared to the traditional empirical approach, providing a useful parameter-selection reference for high-precision wireless charging alignment systems under the investigated static operating conditions.
In response to the significantly reduced efficiency of magnetic coupling wireless charging for unmanned aerial vehicles (UAVs) caused by their high sensitivity to transmitter and receiver coil alignment, as well as landing point errors, a position identification method based on the detection coil-induced voltage fingerprint and embedded neural network regression is proposed. This enables position alignment through a 2D mechanical structure. Firstly, by means of an S–S compensation topology with a bipolar (BP) symmetrical four-detection-coil array deployed at the transmitter, the system effectively suppresses primary direct coupling, ensuring that the position of the receiver coil predominantly determines the detection signals. Secondly, by establishing a voltage fingerprint database during the offline stage and utilizing a multi-layer perceptron–radial basis function (MLP-RBF) regression model, the system achieves high-precision end-to-end positioning and alignment control during the online stage through induced voltage acquisition and data processing. Finally, experiments demonstrate that the proposed method achieves centimeter-level positioning accuracy, with an average error of approximately 1.2 cm and a maximum error of less than 1.8 cm, presenting excellent deployability and engineering applicability.
24 September 2025 in IEEE Transactions on Power Electronics
This article proposes a novel primary side controllable magnetic flux capacitor (CMFC) for a wireless charging system (WCS) in automatic guided vehicles (AGVs). It aims to address two challenges: zero-voltage switching (ZVS) for the full-bridge inverter and constant power (CP) charging for the battery, which are affected by variations in the transmitter coil’s self-inductance and the magnetic coupler’s mutual inductance, as well as the battery’s equivalent load resistance. The CMFC also replaces the resonant capacitor on the primary side. First, the circuit structure of the designed WCS for AGVs and the two problems existing in it are analyzed. Therefore, the CMFC is proposed to address the drawbacks of similar existing methods in resolving these problems. Second, from the hardware aspects, the concept of “controllable magnetic flux principle” and parameter design method for the CMFC is given. From the software aspects, the closed-loop control strategy for the WCS with CMFC is given. Finally, an engineering prototype is constructed. It verifies that the designed WCS can achieve 350-W CP charging and 92.7% efficiency with equivalent load resistance from 2.57 to 8.33 Ω and misalignment up to 40% or 10% of the transmitter coil’s width in x- or y-direction.
To address the residual lateral coil misalignment after an unmanned aerial vehicle (UAV) lands on a fixed wireless-charging platform, this study proposes a graph-attention-based knowledge distillation method for embedded self-alignment localization. Four detection-coil voltages form an induced-voltage fingerprint database organized as a multi-scale spatial graph. A graph attention network (GAT) teacher model is trained offline to learn neighborhood correlations in the voltage–position mapping, and its spatial knowledge is distilled into a lightweight Tiny-MLP student model for microcontroller unit (MCU)-based online inference. Experimental results show that the GAT teacher achieves a mean absolute error (MAE) of 0.589 cm, while the distilled Tiny-MLP reduces the MAE of the directly trained Tiny-MLP from 1.548 cm to 1.148 cm (a 25.8% reduction under a fixed seed). In 2000 closed-loop alignment trials with random initial positions, the system achieves an 85.5% success rate under a 0.5 cm threshold, indicating that the method supports low-complexity closed-loop self-alignment for UAV wireless charging.
Magnetically coupled wireless power transfer is highly sensitive to the relative height between the transmitter and receiver coils. Meanwhile, complex electromagnetic interference and measurement noise induce fluctuations in the voltage fingerprints and degrade the fingerprint-library quality, which in turn leads to an unstable closed-loop alignment. To address these issues, this study proposes a position identification and closed-loop alignment method based on decoupled sensing-coil voltage fingerprints and integrated neural network regression. First, a four-quadrant symmetric decoupled sensing-coil array is deployed at the transmitter, and a compensation network is introduced to suppress direct-coupling effects. Thus, the multi-channel sensing voltages are primarily driven by the planar position variation of the receiver coil, thereby forming voltage fingerprints with improved spatial separability. Second, an original evaluation metric for centimeter-level alignment reliability is developed, and a lightweight residual calibration strategy is designed to target hard samples and tail errors to alleviate the error amplification caused by fingerprint noise and distribution imbalance. On this basis, a multi-model fusion integrated neural network is constructed to enhance disturbance rejection and prediction consistency. Finally, closed-loop alignment experiments are conducted over a two-dimensional workspace and compared with representative baseline methods. The results demonstrate that the proposed method achieves more concentrated terminal alignment positions and more stable closed-loop behavior, even when the fingerprint library is of low quality, thus validating its engineering applicability and robust alignment capability.