Multiband Four-Port MIMO Antenna Design at Terahertz Frequencies Using Machine Learning-Based Optimization
摘要
This chapter presents the performance assessment of a four-port connected MIMO antenna system. It is designed for multiband operation across terahertz frequencies. The antenna features a compact size of 120 μm × 120 μm, utilizing a lossy polyimide substrate. Both the patch and ground layers are made of annealed copper, ensuring efficient conductivity. The antenna resonates at multiple frequencies—1.750 THz, 2.590 THz, 4.480 THz, and 7.410 THz—demonstrating effective multiband performance. At these frequencies, the reflection coefficients (S11) are observed to be −38.17 dB, −24.38 dB, −30.33 dB, and −25.28 dB, respectively. The transmission coefficients (S21) at the same resonant frequencies exhibit similar values of −22.78 dB, −19.74 dB, −26.30 dB, and −20.12 dB, indicating strong isolation between the antenna ports. Furthermore, machine learning algorithms are employed to optimize critical parameters such as S21 and TARC enhancing the overall design efficiency and performance validation. This combination of advanced optimization techniques and compact design makes the antenna suitable for high-frequency terahertz applications.