Component-based Framework for Data Preprocessing within Connected Car Architectures: Insight of future software development and end-to-end simulation

摘要

The ongoing development of vehicle-extended communication networks and cross-fleet data storage is driving an increasing focus on connected car functions. The outsourcing of software enables the usage of large and temporally or locally distributed data sets and algorithms. Considering current infrastructures, the quality of service is limited by mobile radio transmission. In order to avoid overload or failure in remote data communication systems preprocessing of car data plays a key role. The methods for sensor data processing are sufficiently investigated, as there are numerous algorithms and derivatives for various identified forms of problems. Based on this the arrangement and location of functionalities have to be investigated. Beginning by preliminary discussing future architectures for connected cars, corresponding software modularity as well as in- and output interfaces are defined for a reasonable structure of basic services. Furthermore, a framework for data preprocessing is presented by identifying generic tasks. Due to the possibility of arbitrary arrangement in a derivative service, functional restrictions and dependencies are carried out. Finally, the findings of structuration are given by simulating possible localizations of the divisible service.