A Generalized Branch Line Coupler with An Arbitrary Power Ratio and Phase Difference Using Artificial Neural Network Optimization

This article presents the design methodology of an innovative generalized branch line coupler using Artificial Intelligence and Machine Learning tools. Unlike the conventional branch line coupler, the proposed generalized four-port coupler can be realized with an arbitrary power ratio and arbitrary phase difference between the output ports, while maintaining a matched condition at all the ports and achieving the required isolation. As a validation of the proposed concept, various generalized couplers with equal and unequal power ratios and arbitrary phase differences are designed. These designs are validated in circuit schematics and 3-dimensional electromagnetic simulators. Furthermore, a design table is also presented in this article for equal power division with arbitrary phase difference (from 45 and 135 degrees, considering practical implementation). In addition, the generalized couplers are also designed using lumped elements along with the microstrip technology. Such generalized couplers can lead to significant performance enhancement through system miniaturization in RF and Microwave transceiver components, especially in beam-forming networks, like Butler Matrix and Nolen Matrix.