Random forest in design of surface acoustic wave resonators based on multilayer structures

Abstract Surface acoustic wave (SAW) resonators are essential to communication technologies, particularly in radiofrequency front-end modules, which require optimized resonator designs to improve their performance. Traditional design methods, such as the finite element method and coupling of modes (COM) theory, are often limited owing to computational inefficiency and associated parameters. We introduce a novel method that leverages the random forest algorithm to predict the admittance with enhanced accuracy and efficiency. The random forest algorithm integrated with feature extraction is adopted. This framework not only streamlines optimization by facilitating the evaluation and selection of structural parameters but also demonstrates versatility in predicting the admittance of different SAW resonators based on silicon carbide. Experimental validation further confirms the accuracy and robustness of the proposed method.