RBFNN-Based Ultra-Wideband Super-Miniaturized $$4\times 4$$ Highly-Isolated MIMO Antenna for 5G mm-Wave Wireless Communications

Abstract

As the demand for high-speed communication in 5G increases, this article presents a development breakthrough in millimeter-wave spectrum solutions. The focus is on designing an ultra-wideband (UWB) circular patch antenna for 5G millimeter-wave applications. To address the problem of limited bandwidth and efficiency in existing designs, a UWB antenna operating in the 20–45 GHz range with a compact form factor of \(4 \times 4.8 \times 0.508 \,\,{\hbox {mm}}^{3}\) has been developed. A radial basis function neural network (RBFNN) was employed to analyze and optimize the antenna’s bandwidth characteristics. The resulting antenna exhibits a wide bandwidth of 25 GHz, a gain of 5.75 dB, and an impressive efficiency of 99%. To extend the capabilities of the design, a \(4 \times 4\) MIMO antenna system was developed, incorporating four copies of the proposed UWB single antenna, achieving an extended bandwidth of 25 GHz. The MIMO system, with dimensions of \(17.5 \times 17.5 \times 0.508\,\,{\hbox {mm}}^{3}\), demonstrates excellent performance with isolation exceeding 30 dB, a gain of 6 dB, and high efficiency. Rigorous measurements validate the designs, affirming their practical viability for future 5G mmWave applications. In conclusion, the proposed UWB antenna and MIMO system offer significant advancements in 5G mmWave communication, providing high performance in bandwidth, gain, and efficiency, supported by comprehensive simulations and measurements.