Architectural Design and Verification/Validation of Systems with Machine Learning Components

Purpose and goal

Machine Learning components requires specific architectural design and verification/validation methods. Architectures containing both stochastic and traditional deterministic components shall be defined which assures safety, robustness and fault tolerance. Furthermore ML algorithms require specific verification and validation methods. The project aims to: * Define and evaluate architectural styles and patterns for developing systems with ML components. * Develop methods for validation and verification, enabling use of ML for safety relevant vehicle systems.

Expected results and effects

In the short term, the project is expected to contribute to the development of safe, robust and fault tolerant electrical systems with ML components in vehicles. In the long run, the project is expected to contribute to the continuous deployment of ML components to different autonomous subsystems in cars and their safe and reliable utilization in the transportation eco-system.

Planned approach and implementation

Based on a technical report on the current state-of-the-art in the areas of architectural design and V&V of systems with ML components we aim to deliver: * Methods and high level requirements for designing automotive system architectures with ML components. * Definition of the set of quality indicators and methods to infer them from ML models and data. * Development of the reference architecture for the future automotive systems that include ML components. * Development of the tool chain architecture for V&V of ML components and models/algorithms.