AI for Automated Modular Production

To stay competitive in dynamic markets, manufacturing companies must adapt flexibly to changing production conditions and objectives. Modular manufacturing concepts enable the production process to be divided into individual, automated production steps without manual programming or configuration.

The European research project RICAIP (Research and Innovation Centre on Advanced Industrial Production) introduces a fully functional demonstrator that shows the fundamental principles of flexible production using a miniaturized production cell. The cooperative system, which consists of two KUKA robots – one stationary and one mobile – autonomously dismantles various types of electric batteries, replaces specific components, and assembles still-usable battery modules into new units. These can be repurposed as large stationary batteries for households or businesses, for example.

The demonstrator combines various AI technologies, such as machine learning and computer vision, with collaborative robotics, network communication, edge computing, and classic automation technology. The centerpiece is the automatic planner, developed by project partner CIIRC at the Czech Technical University in Prague. The planning software compares available resources with the machine functions and independently creates the production sequence – without manual programming or configuration. Moreover, the planner is integrated in an industrial Manufacturing Execution System developed and operated by T-Mobile.

The modular structure of the robotic production cell, the seamless integration of individual production machines, and the dynamic, machine-based rescheduling of production steps make this concept unique. The demonstrator serves as proof of concept for the automated retrofitting of multi-component products and can be transferred to different product types.