Domain Decomposition Based Artificial Neural Networks (ANNs) Modeling of Acoustic Wave Resonators and Filters

This paper investigates the effectiveness of artificial neural networks (ANNs) as a surrogate modeling method based on machine learning algorithms in emulating the electroacoustic wave behavior of the high-Q piezoelectric resonators and filters. A domain decomposition approach incorporating ANN models to concurrently analyze multidomain radio frequency (RF) modules is also discussed. Different multilayer perceptron (MLP) ANN models have been developed and benchmarked against their model accuracy and model efficiency. The developed models are then utilized to construct ladder-type Band 7 and Band 41 bandpass transmit filters, as examples, to highlight the quality of the modeling method. Other possible applications pertinent to the capability of machine learning algorithms are briefly discussed.