研究目的
To propose a better solution for miniaturized antenna design using a circular fractal geometry and a hybrid soft computing approach combining BFO, PSO, and ANN ensemble to achieve accurate design for 2.45 GHz ISM band applications.
研究成果
The proposed hybrid algorithm successfully designs a miniaturized CFP antenna with a 35.5% size reduction at 2.45 GHz, validated by good agreement between simulated and experimental results. The method offers improved accuracy, convergence, and generalization for fractal antenna design, with potential for extension to other shapes and frequencies.
研究不足
The design relies on assumptions to reduce the number of independent variables, which may limit generalizability. Fabrication errors can cause deviations between simulated and measured results. The hybrid algorithm may be computationally intensive, and the approach is specific to the chosen antenna geometry and frequency band.
1：Experimental Design and Method Selection:
The methodology involves designing a circular fractal patch (CFP) antenna and developing a hybrid algorithm that integrates a PSO-based selective ANN ensemble with a BFO algorithm. The ANN ensemble models the resonant frequency based on the antenna's outer radius, and this model is used as the objective function in BFO to optimize the radius for a target frequency of 2.45 GHz.
2：45 GHz.
Sample Selection and Data Sources:
2. Sample Selection and Data Sources: Data for training and testing the ANN ensemble are generated using IE3D electromagnetic simulator. Forty random values of the outer radius R between 3 and 15 mm are simulated to obtain corresponding resonant frequencies, with 31 pairs for training and 9 for testing.
3：List of Experimental Equipment and Materials:
The antenna is fabricated using FR4 substrate (dielectric constant εr = 4.3, thickness h = 1.47 mm) and coaxial probe feed. Measurement is done using a vector network analyzer (Model ZVL).
4：3, thickness h = 47 mm) and coaxial probe feed. Measurement is done using a vector network analyzer (Model ZVL).
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: Steps include: designing the CFP antenna geometry, generating simulation data with IE3D, training the ANN ensemble with bagging and PSO for weight optimization, hybridizing with BFO to find optimal R, simulating the optimized antenna, fabricating a prototype, and measuring S11 and radiation patterns.
5：Data Analysis Methods:
Performance is evaluated by comparing simulated and measured resonant frequencies and radiation patterns, using error metrics and visual analysis.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
