Self-validation of complex electronic systems in safety-critical mobility applications based on greybox models

Project Description

AI-based enhancement of the reliability of electronics for (autonomous) mobility

PROJECT GOALS

The requirements for electronics in safety-critical applications regarding functional safety and availability are very high. In particular, the areas of mobility offer extensive potential hazards in case of malfunction, manipulation, or failure. Typical application areas include monitoring systems for train control and control units of electrified vehicles with autonomous driving functions, which are to be considered within SesiM. In addition to individual central components, electronic assemblies themselves offer high innovation potential to guarantee functional integrity. The main goal of the collaborative project is to develop a hybrid, model-based condition monitoring of complex electronic and mechatronic systems and to prototype implementation in relevant applications of automotive and railway technology, e.g., safety-relevant electronic systems for train control and control units of electrified vehicles with autonomous driving functions.

INNOVATION & METHODOLOGY

A central aspect of SesiM is the development of an AI-based condition monitoring for the optimized operation of automotive and railway technology. A digital fingerprint of the electro- and mechatronic assemblies is generated to proactively respond to aging-related wear and safety-critical changes. Changing influences from manufacturing processes and material qualities, ex- and intrinsic stresses during the usage phase, as well as system-descriptive sensor data are collected, evaluated, and utilized within an innovative model formation. The novel approach is integrated into a self-diagnosis at the system level and an intelligent operation and maintenance management is realized. In safety-critical systems, considerable over-engineering or redundant structures are currently used to avoid failures.

OUR CONTRIBUTION

The core of the project contribution by Gestalt Robotics is the development of tailored AI methods for the inspection of electronic components as well as the development of a cross-cutting communication architecture for the self-validation of electronic components. Regarding the implementation of common pattern applications and demonstrators, participation in the concept creation for sensors, online monitoring, and data generation is carried out with consideration of the application-specific requirements regarding data augmentation, generation of synthetic image data, as well as active learning approaches. In this context, the suitability of few-shot learning methods will also be tested. For the communication architecture, data protection and ownership aspects regarding image data are additionally examined.