An Application of Shape Similarity Recognition Using PCA based Dimensional Compression

In recent years, the demand for faster product development has been increasing year after year. In addition, as requirements and their levels become more sophisticated, data-driven development that makes use of past data is attracting attention. Compared to experiments, simulations are characterized by the ease of retaining data that can be used for analysis, but this creates the problem of handling huge amounts of result data. In order to overcome this challenge, we propose a method to detect similar behaviors based on the distance in the modal space obtained from the animation results of past calculations with a reduced dimension reduction technique.