Fast Masked Quantum Tyrannosaurus Convolutional Attention Network Optimized Dual‐Port Dual‐Band 28/34 GHz MIMO Microstrip Antenna for High‐Efficiency Satellite Communication

Abstract The increasing growth of satellite communications requires very efficient, dual‐band millimeter‐wave MIMO antennas across frequency bands. Dual‐port, dual‐band microstrip patch MIMO antenna design operating at 28 and 34 GHz overcomes the issues of limited substrate area, mutual coupling, and frequency selectivity. Using the Fast Masked Quantum Tyrannosaurus Convolutional Attention Network (Fast‐MQT‐CAN), a new combination of Fast Quantum Convolutional Neural Networks with Masked Attention Transformers, and optimized using the Tyrannosaurus (T‐Rex) Optimization Algorithm (TROA), the antenna is able to have accurate dimensioning and optimal positioning of patches over a small substrate. The proposed design has a radiation efficiency of 90%, which is high, allowing for excellent signal strength and low energy loss. Other advantages are broadband operation with constant gain, and with no significant inter‐element interference, which would be useful for future satellite communication applications.