Machine learning inspired compact quad port MIMO antenna with circular polarization for multi band wireless applications

In this work, a compact quad-element MIMO antenna demonstrating resonance at 6.39 GHz, 11.09 GHz, 14.69 GHz, 17.96 GHz encompassing C, X, and Ku bands, respectively within a single compact structure. The designed quad-element MIMO antenna employs a Y-shaped radiating monopole configuration with connected ground elements for easy system integration. Also, the designed MIMO antenna attains geometry driven Circular Polarization (CP) at specific elevation angles without any external phase-shifters/ parasitic structures, resulting in enhanced robustness against multipath fading and polarization mismatch. Further, the critical requirement of high inter-element isolation is maintained (Sij < - 20 dB), suppressing mutual coupling effectively. The parameters of the presented quad port antenna design are also optimized using various machine learning (ML) algorithms and Gaussian regression process model is found to be most suitable providing best return loss performance. Finally, the designed quad port MIMO antenna is fabricated, and its prototype is tested to validate the simulated results. Further, the comprehensive analysis of diversity performance demonstrates ECC (≤ 0.001), DG (≈10 dB), CCL (< 0.2 bps/Hz), and TARC (> 9.99) parameters yielding an ergodic channel capacity exceeding 9 bps/Hz. The average efficiency is observed to be of 57% across all bands with a peak gain of 9.78 dBi at 17.96 GHz achieving highly integrated, reliable, promising solution for CP-MIMO communication systems making it a strong candidate for satellite and advanced wireless applications.