Table of Content

15 July 2026 Volume 39 Issue 7

  

A Hybrid Control Method with Linear Active Disturbance Rejection for WPT System

DING Hang, KAN Jiarong, XU Jiajun, JIN Wei, ZHENG Yang

Electronic Science and Technology. 2026, 39(7):  1-6.  doi:10.16180/j.cnki.issn1007-7820.2026.07.001

Abstract ( 58 )   HTML ( 1 )   PDF (5093KB) ( 31 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems of unstable power transmission and low efficiency of the wireless power transmission system when it is affected by external interference, this study proposes a composite control method for wireless power transmission with high efficiency and high stability based on linear active disturbance rejection control. The effective values of the voltage and current at the input end of the wireless power transmission system are sampled to identify the coupling coefficient. The impedance matching is carried out by adjusting the duty cycle of the Buck-Boost circuit on the output side to achieve maximum efficiency tracking. The linear active disturbance rejection control is adopted to adjust the phase shift angle of the full-bridge inverter on the input side to ensure the stable and efficient power transmission of the system. The proposed method realizes the decoupled operation between the two control loops of high-efficiency power transmission and constant voltage output of the wireless power transmission system. The simulation results show that the system can stably transmit power when affected by external interference, and the system transmission efficiency is maintained above 90%. Compared with the proportional-integral control, the linear active disturbance rejection control has a better dynamic regulation effect on the output voltage when facing load disturbances and reference voltage disturbances.

Research of Leaf Properties Measurement Algorithm Based on Image Processing

GUAN Jing, LOU Fei

Electronic Science and Technology. 2026, 39(7):  7-13.  doi:10.16180/j.cnki.issn1007-7820.2026.07.002

Abstract ( 57 )   HTML ( 3 )   PDF (1411KB) ( 21 )  

Figures and Tables | References | Related Articles | Metrics

Plant leaves are the main organs for photosynthesis and transpiration in plants. The growth environment and health condition of plants are related to the size of their leaves. Plants can be classified based on the characteristic attributes of their leaves. In order to calculate the area and circumference of plant leaves, a leaf attribute measurement algorithm based on image processing is proposed. The plant leaf images are preprocessed such as enhancement and filtering to improve the image quality and reduce the influence of noise. The leaf edge contour and leaf target area are extracted after image segmentation using the maximum between-to-class variance method. To address the issue of hollow points in the target area, morphological operations are used to remove the hollow points and extract the entire area of the leaf. The number of pixels is counted to obtain the leaf area. The perimeter attribute is calculated based on the blade profile and chain code direction code to improve the accuracy of perimeter measurement. The experimental results show that the proposed algorithm can accurately measure the attributes of plant leaves, has high measurement accuracy, and is applicable to different plant species.

A Fusion of Residual-Driving and Cluster-Adaptive Intuitionistic Fuzzy C-Means Image Segmentation Algorithm

ZHANG Hao, SONG Yan, DOU Jun

Electronic Science and Technology. 2026, 39(7):  14-23.  doi:10.16180/j.cnki.issn1007-7820.2026.07.003

Abstract ( 35 )   HTML ( 1 )   PDF (5868KB) ( 10 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems that the existing intuitionistic fuzzy clustering algorithms have difficulty in determining the optimal number of clustering clusters and are affected by noise points during the clustering process, an IFCM(Intuitionistic Fuzzy C-means) image segmentation algorithm integrating residual-driven and cluster adaptive is proposed. To alleviate the fuzziness and uncertainty of sample points, based on the intuitionistic fuzzy set framework, both sample membership degree and hesitation degree are considered simultaneously, thereby obtaining a more accurate membership degree. Based on the traditional IFCM, a regularization term of cluster adaptive merging is introduced, enabling the number of clusters to be adaptively adjusted and the optimal number of clusters to be found. This effectively avoids the cumbersome setting of the initial number of clusters and reduces the sensitivity of the initialization parameters. The effectiveness of the proposed algorithm is verified on the artificial image dataset and the magnetic resonance image dataset. The experimental results show that the proposed algorithm can achieve a better noise removal effect while adaptively determining the optimal number of clusters.

Feature Selection Method Based on Kernel Density Peak Clustering and Fine-Grained Noise Suppression

LIANG Runchen, SONG Yan, DOU Jun

Electronic Science and Technology. 2026, 39(7):  24-32.  doi:10.16180/j.cnki.issn1007-7820.2026.07.004

Abstract ( 34 )   HTML ( 1 )   PDF (5378KB) ( 6 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems that traditional feature selection methods are vulnerable to noise and the obtained feature space is prone to cause changes in data distribution, this study proposes a FNKC(Feature Selection based on Fine-Grained Noise Suppression and Kernel Density Peak Clustering). To overcome the influence of noise on feature selection, the possibility theorem of the PFCM (Possibilistic Fuzzy C-means Clustering Algorithm) is utilized in combination with the information particle criterion to extract data correlation, thereby improving the anti-noise performance. Introducing kernel density in high-dimensional space to measure clustering density can accurately capture the spatial structure within clusters, thereby better reflecting the distribution of data. Finally, the superiority and effectiveness of the proposed algorithm are verified by comparing multiple advanced feature selection methods on six public high-dimensional datasets.

A Keyframe-Based Approach for Facial Expression Mimicry in Humanoid Robots

LI Jiahao, YU Qi, YUAN Ye

Electronic Science and Technology. 2026, 39(7):  33-40.  doi:10.16180/j.cnki.issn1007-7820.2026.07.005

Abstract ( 48 )   HTML ( 2 )   PDF (6579KB) ( 13 )  

Figures and Tables | References | Related Articles | Metrics

To enhance the expression imitation ability of humanoid robots in human-computer interaction, an expression imitation strategy based on facial keyframe detection and key point recognition is proposed. By establishing a keyframe detection network through self-supervised multi-scale optical flow technology, it is possible to effectively capture subtle changes in facial expressions and identify keyframes in expression sequences. The facial key point detection algorithm is adopted to extract the expression data that the robot needs to imitate, and this strategy is deployed on the hardware platform of the humanoid robot. Experiments are conducted on public datasets such as Vox, 300VW and CAER. The results show that the average accuracy rates of keyframe detection are ±3.61 and ±0.75 respectively, and the cosine similarity between the expressions imitated by the robot and the real human expressions is 0.722 6. The experimental results show that the proposed strategy significantly improves the naturalness and fluency of robot expression imitation, which proves an effective solution for humanoid robots to imitate expressions in human-computer interaction and indicates great potential in practical applications.

A 20~40 GHz Broadband High-Precision Six-Bit Digitally Controlled Phase Shifter

ZHANG Bin, QIN Zhanming, WANG Baikang, JIANG Yingdan

Electronic Science and Technology. 2026, 39(7):  41-47.  doi:10.16180/j.cnki.issn1007-7820.2026.07.006

Abstract ( 37 )   HTML ( 1 )   PDF (7064KB) ( 7 )  

Figures and Tables | References | Related Articles | Metrics

In order to meet the high-precision design requirements of Ka-band wireless communication systems, a 20-40 GHz wideband high-precision six-bit digitally controlled phase shifter is designed in this study using the GaAs 0.15 μm PHEMT(Pseudomorphic High Electron Mobility Transistor) process. By cascading six different phase shifts, a phase shift range of 0° to 360° with a minimum step of 5.625° can be achieved. Among them, the 5.625° phase shift adopts a symmetrical CLC embedded switch linear structure, while the 11.25° and 22.50° phase shifts adopt a bridge T-type topology. The 45° and 90° phase shifts adopt the Lange coupler-type reflective structure, while the 180° phase shift adopts the switch-selective structure based on the Lange coupler. The electromagnetic simulation results show that within the frequency band range of 20~40 GHz, the 64-state phase RMS(Root Mean Square) error of the phase shifter is less than or equal to 5°, both the input and output return losses of the port are greater than 10 dB, the insertion loss is less than 10 dB, and the 64-state insertion loss fluctuation of the phase shifter is less than ±0.6 dB.

Study on Model Predictive Control Strategy Based on Adaptive Moving Average

DENG Lihong, YANG Chao

Electronic Science and Technology. 2026, 39(7):  48-55.  doi:10.16180/j.cnki.issn1007-7820.2026.07.007

Abstract ( 39 )   HTML ( 2 )   PDF (3402KB) ( 11 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problem of charge and discharge imbalance caused by efficiency in the actual operation of the adaptive moving average algorithm's suppression results, this study proposes a control strategy based on model predictive control to optimize and adjust the target power of energy storage. Based on the adaptive window moving average algorithm for solving the energy storage power, a hybrid energy storage system composed of lithium-ion battery and supercapacitors is adopted, and the complementary set empirical mode decomposition method is combined to obtain the energy storage target power and the preliminary energy storage capacity. Based on the operational characteristics of the wind power generation system and the demand for fluctuation suppression, the state space equation and related constraint conditions are established for model predictive control to optimize the energy storage power. Through the simulated annealing algorithm, with the goal of minimizing the full life cycle cost of energy storage, the optimal capacity of the hybrid energy storage system is calculated. The case study analysis shows that in the hybrid energy storage system with the same capacity, the model predictive control strategy achieves energy balance in the hybrid energy storage system and optimizes the charging and discharging power of the energy storage. In addition, the hybrid energy storage capacity calculated by the simulated annealing algorithm further reduces the life cycle cost.

Robust Collaborative Scheduling Strategy of Multi-Microgrid Based on Non-Cooperative Game and Income Distribution Mechanism

FU Taotao, YANG Chao

Electronic Science and Technology. 2026, 39(7):  56-66.  doi:10.16180/j.cnki.issn1007-7820.2026.07.008

Abstract ( 36 )   HTML ( 1 )   PDF (10720KB) ( 22 )  

Figures and Tables | References | Related Articles | Metrics

The uncertainty of renewable energy output and fluctuations in electricity prices affect the management of electricity sharing. To determine the fairness of power sharing payment in multi-microgrid systems, this study constructs a two-stage robust collaborative scheduling model for multi-microgrids based on non-cooperative games and revenue distribution mechanisms, thereby reducing the total operating cost of multi-microgrid systems and ensuring that each microgrid can benefit from the proposed model. In the first stage, the fluctuation of electricity prices and the uncertainty of wind and photovoltaic output are considered, and a multi-microgrid power management model is established using robust processing to achieve point-to-point power sharing among the microgrids. In the second stage, a CRRD (Cost Reduction Ratio Distribution) model based on non-cooperative game is established for the revenue distribution mechanism of shared electric energy, and the generalized Nash equilibrium is used to determine the clearing and settlement of shared electric energy. The first stage problem is solved using the C&CG(Column and Constraint Generation) and the second stage problem is solved using the ADMM(Alternating Direction Method of Multipliers). The results of the case study analysis show that the proposed model can reduce the operating costs of each microgrid.

Capsule Network Based on Mask and Attention

ZHOU Junwei, SONG Yan, ZENG Ru

Electronic Science and Technology. 2026, 39(7):  67-75.  doi:10.16180/j.cnki.issn1007-7820.2026.07.009

Abstract ( 40 )   HTML ( 1 )   PDF (4960KB) ( 7 )  

Figures and Tables | References | Related Articles | Metrics

In view of the problems of poor feature extraction ability and easy formation of capsule redundancy in traditional capsule networks, a capsule network based on mask and attention is proposed. In view of the problem that the combination of masks may cause insufficient information in some capsules, this study proposes a capsule separation operation to effectively separate the capsules with less information and reduce their weights. To enhance the feature extraction capability of the capsule network, a mask is integrated into the feature extraction layer, and capsule separation is adopted to solve the problem of insufficient capsule information caused by the mask. Adding capsule attention to the capsule layer enables the model to focus on more important capsules, effectively avoiding capsule redundancy. The experimental results show that the classification accuracy of the proposed model on the MNIST and SVHN datasets is 99.77% and 98.33% respectively, and on the CIFAR10 and F-MNIST datasets is 94.81% and 95.46% respectively. When facing datasets with affine transformation and adversarial sample attacks, the proposed model can still maintain a high accuracy rate, fully demonstrating its good robustness.

Research Progress of High Precision Additive Manufacturing Control Algorithms

YANG Yunhui

Electronic Science and Technology. 2026, 39(7):  76-80.  doi:10.16180/j.cnki.issn1007-7820.2026.07.010

Abstract ( 28 )   HTML ( 1 )   PDF (2001KB) ( 5 )  

Figures and Tables | References | Related Articles | Metrics

To enhance the positional accuracy of additive manufacturing in the three-dimensional discrete stacking process, an increasing number of high-precision additive manufacturing control algorithms have been applied and studied by the domestic and international industrial sectors. Path planning in additive manufacturing requires the use of multiple high-precision control algorithms to address issues such as step effects, idle paths, and motion position deviations, while enhancing manufacturing efficiency and product quality. This study sorts out and reviews the main problems causing accuracy errors in additive manufacturing, discusses the current main filling algorithms and analyzes the advantages, disadvantages and applicable conditions of the filling algorithms, and summarizes the common genetic algorithms and ant colony algorithm principles for dealing with idle paths, as well as the research progress of adaptive control systems in additive manufacturing. The problems that additive manufacturing has not been widely applied on a large scale and the key issues that need to be addressed in the future have been proposed..

License Plate Detection and Recognition System Based on Differential Cameras in Strong Backlight Conditions

HE Guotao, YU Boyu, LI Xianrui, BAI Wanjing, WANG Bo

Electronic Science and Technology. 2026, 39(7):  81-90.  doi:10.10980/j.cnki.issn1007-7820.2026.07.011

Abstract ( 34 )   HTML ( 3 )   PDF (8121KB) ( 5 )  

Figures and Tables | References | Related Articles | Metrics

In view of the the problems of the decline in license plate recognition performance caused by the limited dynamic range of traditional frame-based cameras under strong backlight, an end-to-end license plate recognition method based on dynamic event cameras is proposed. Dynamic event cameras have the advantages of a dynamic range greater than 126 dB, only outpacing brightness changes, and overcoming the failure of frame-based imaging. The event stream of license plates with strong backlight is collected by using the self-developed camera. The number of events is mapped to grayscale images through the cumulative sliding time window to construct the first event dataset DVS-PD(Dynamic Vision Sensor Plate Dataset) for license plate recognition with strong backlight. The recognition network adopts the STN(Spatial Transformer Network) to correct the tilted license plate, and the back end uses the LPRNet(License Plate Recognition Network)to output the character sequence. In the training section, the CTC(Connectionist Temporal Classification)loss and data augmentation are introduced, and the two denoising preprocessing methods of STCF(Spatio-Temporal Correlation Filter) and BES(BackgroundEvent Suppression) are compared. The experimental results show that the recognition accuracy rates of the proposed model in the three types of event frames, namely the original, STCF and BES, are 93.4%, 94.5% and 95.1% respectively, which are superior to the LPDRNet(License Plate Detection and Recognition Network), the RCNN(Recurrent Convolutional Neural Network) and the SLPNet(Sequence License Plate Recognition Network). It has been proved that the dynamic event camera can effectively solve the problem of license plate recognition in strong backlight. The constructed dataset provides a benchmark for subsequent research. The proposed model combines lightweight and high inference speed, and can effectively meet the scene requirements of real-time license plate detection and recognition.