Integration and safeguarding of the overall AI system
To enable agile development, system and software architectures are being developed for four test vehicles and a central cloud-based data processing architecture is being created.
The work package defines the integration architecture for all vehicles, integrates the systems and supports the execution of the planned tasks for the vehicles set up in KARLI. In addition, a central cloud-based data processing architecture will be created and optimized in terms of performance and cost. For the KI-FFZ and KI-MMI, application-specific safeguarding methods will be developed to ensure AI robustness and avoid misuse and malicious attacks.
Consequently, work package 600 deals with the integration of the necessary sensor technology of the different applications into the test vehicles as well as the connection to a centralized data infrastructure according to the use cases and the methodical safeguarding of AI functions.
To ensure smooth integration in the creation of the interface architecture, all involved partners come together to jointly develop the interfaces based on the requirements.
This acts as the final step in the development of the applications and represents the setup in the test vehicles according to the prescriptions of the interface architecture and the requirements determined in advance. With the help of AI assurance, functional, security and legal assurance are highlighted using concepts and methods.
The aim of the work package is to develop the system and software architectures for four test vehicles. The data collection systems and applications of the four test vehicles are to be integrated in accordance with the use cases, and security concepts are to be designed and implemented.
Ultimately, the goal is to enable agile development by ensuring high availability of vehicles and the required data in the cloud environment, as well as to develop usage guidelines for optimal cloud use.