Transitional Structures In Three-dimensional Microwave Integrated Circuits, Artificial Neural Network Modeling

Advances in developing transitions in microwave integrated circuits during the last decade are reviewed. Some typical structures of transition are introduced. Transition structures can be classified into two basic types: one is transition between the same kind of transmission lines on different planes of a common substrate, the other transition between different types of transmission lines. A new transition structure (microstrip-slotline transition) is presented by analyzing various kinds transition structures. Planar lumped-element parallel LC circuits are used to replace λ/4 transformer structures. The transition has the advantage of small size and wide bandwidth. The transition structure is optimized and simulated by commercial software HFSS and CST Micrwave Studio. In order to validate the simulated results, a back-to-back microtrip-slotline transition structure is designed and fabricated. It is shown that both simulation and measurement are in good agreement. The bandwidth for 1.3dB insertion loss is from 4GHz to 10GHz. For today's high performance and fast electromagnetic engineering computer-aided design, it is increasingly necessary to model elements and devices with high accuracy, reliability and efficiency. Artificial neural network (ANN) recently has received extensive attention as a fast and accuate modeling tool in electromagnetic engineering. Based on the error backpropagation algorithm, multiple layer perceptron artifical neural networks (ANN) are built for the transition structure. The ANN model not only preserves the accuracy of the EM simulations, but also simplifies its CPU and memory requirements, and at the same time keeps good extrapolation capability. Hence, the model has potential power for the high performance and fast computer-aided design (CAD) in electromagnetic engineering.