2024 VI-GRADE ZERO PROTOTYPES SUMMIT

As full vehicle simulation requisites are still growing on the auto industry and more decisions are being made based on simulator evaluations, customers are starting to require complex damper models during RFQ process. Enhancing model techniques is a must in such a dynamic market and complex simulation can be simplified using regression methods (lookup tables), or 1D physical models, while others (multiple input systems) need more robust tools (deep learning methods). These deep learning techniques not only aim for data processing, but also allow to generate active models with years of knowledge embedded into them. By creating an efficient data measurement and data processing procedures, an accurate and fast in-house deep learning model damper models can be created for full vehicle simulations. The presentation will present how Hitachi evolved on these deep learning models, by showing the adhesion of these models with measurement data and their influence in full vehicle simulation environment.