Segmentation of Multivariate Time Series with Convolutional Neural Networks

This paper addresses an important problem of time series analysis at test benches. A new method is presented that allows automated segmentation of measurement time series using a Convolutional Neural Network (CNN). The CNN is trained for this purpose with specifically generated data. The results show a high quality and efficiency. The field of application of the algorithm is not limited to the automotive industry focused on here, but can be easily transferred to other areas that allow a visual segmentation of data.