Machine learning algorithms for impact localization on formed piezo metal composites

Smart materials are of growing interest in research as well as industry due to its wide range of possible applications. Combined with a mass production enabled manufacturing of these materials, its application is further increased. A current research project addresses the integration of piezo ceramic elements into metal sheets to develop a smart touch-alike input device. The aim is to develop a low-cost embedded processing system that is able to compute the origin of impacts affecting the material surface. In this paper, the usability of machine learning algorithms to be used for adaptive impact detection and localization is evaluated. Data generated by suitable data pre-processing techniques is applied to neural networks, support vector machines and extreme learning machines for evaluation.