Search for: All records

Abstract Programmable metamaterials have seen increased interest in recent years due to their suitability for a wide range of applications along with the high level of control that they offer over their structural properties. In particular, significant interest has been generated for their application in vibration control as they can be updated and tuned without having to completely rebuild an entire section. However, their complex behavior can make modelling them difficult and time consuming. In recent years, machine learning has emerged as a powerful tool to predict behavior of metamaterials and help reduce the amount of testing time. While prior work demonstrated the efficacy of machine learning on metamaterials with fewer permutations, minimal focus has been placed on its applications for large datasets. This work aims to bridge this gap and demonstrate the possibility of using machine learning algorithms to predict complex metamaterial behavior which is trained on a relatively small dataset. Discussion is also given to a novel approach to collecting experimental data for similar applications.