Study on In-Vehicle CAN dynamic Service Platform Based on Universal Middleware

Recently, there has been an increase in the importance of building a dynamic service platform to converge the various data generated in the in-vehicle system into a collection method and an external system.This dynamic service platform serves as a vehicle mobility terminal as a merged middleware layer between the control and data modules.In this vehicle mobility terminal, the in-vehicle component constituting the cross-service was fused with the dynamic service as a data module connected to the brake and engine, and the external system was modularized and designed.A study of these modularized services for external systems linked to in-vehicle vehicle diagnosis, management services, and service expansion modularized by dynamic services was presented.Currently, the modular service is created, deleted, updated, driven, and stopped when converting the data and service modules.It is also applied in real time to the framework.First, the personalized service distribution model design required for the vehicle convergence service terminal was applied.Second, we presented a user-centered CAN-based TDCSI(Thin Device Customer Service Interface) for the minimum resource use of embedded systems.Third, a UDP-based generative web service optimization model was constructed.These module configurations interact with communication service and data modules inside the vehicle in one direction, two directions, and user demand directions, respectively, so scenarios can be reflected as a life cycle process according to changes in external service requirements.The life cycle model is linked to various application services transmitted between vehicle external systems for data generated inside the vehicle.This structure that can be recycled in an ultra-large deep learning data platform because it can be classified by class, type, and situation.It is expected to contribute to its role as a sustainable mobility platform by applying deep learning that can be packaged with an external multimodal data analysis field and an ML (Markup Language) model.