Assembly process optimization

 

AUDI, INTRA

Based on computer vision systems and deep learning methods, this demonstrator ensures correct installation and enables an ergonomic evaluation of the workers' activities. The functions of the AI systems are described by the following data collection regarding the motion of workers and as-built situation, response per car if it is built as specified and which errors are detected, response per shift regarding the movement of workers, and which level of ergonomics they reached (anonymised data).

Beyond state-of-the-art developments and impacts


A platform module is implemented with the solution, which will enable computer vision systems to be used in different processes, independent of parts. Integrating the digital factory models into the learning algorithms will also be possible to learn new processes quickly and efficiently. The connection of different source and target systems is possible through Docker interfaces. Additional information for the ergonomic evaluation of the work is used to improve and secure the work system and the problem-solving process.