Design of SAW based MEMS device for self-powered electronics using artificial intelligence technique

Piezoelectric microelectromechanical system (MEMS) converts mechanical motion to electrical energy for low-power application. The device can be used continuously to harvest energy from the environment and therefore the electrical energy generated is clean and renewable. Due to this, piezoelectric MEMS is increasingly favorable and stands as potential candidate for self-powered Internet of Thing (IoT) based electronic products and wireless electronic systems. Conventionally, piezoelectric devices used for energy harvesting are based on cantilever structure which is generally fabricated through series of dedicated manufacturing processes. Surface acoustic wave (SAW) based piezoelectric device on the other hand can be fabricated via monolithic photolithography technique which is much simpler and cost effective compared to the fabrication of the cantilever in micro/nano scale. However, the design procedure of such SAW device for self-powered electronic is generally much complex and complicated to be followed than the earlier design. Hence, this paper presents the exploratory work in designing the SAW based MEMS device for self-powered electronic using customized back propagated deep learning neural network. The limitations of the work and related resolutions are highlighted and discussed in the paper. Overall, the result shows promising outlook in using the artificial intelligence technique for design automation of the SAW device.