DiProLeA

Across multiple branches, the product development process is divided into three phases: concept development and production. Within the companies, product-specific information from these phases is still stored in various systems and databases for specific tasks. The networking of these systems and databases offers a high potential to meet the increasing variety of variants and the increasing quality of requirements in manufacturing in the future. Together with the partners involved in the process, a modular and industry-independent socio-technical assistance system for the product development process is to be developed.

Aschaffenburg UAS contributes research and concepts to the design of the assembly workplace of tomorrow. This is of particular interest because manual assembly processes performed by humans will remain indispensable even in the digital factory of the future – whether because of the high flexibility or the high precision that this ensures. It is aspirational to intelligently support humans in these mostly monotonous but nevertheless concentration-demanding activities. Existing assistance systems for manual assembly processes are usually elaborately optimized for individual processes. These systems are not capable of generating dynamic assistance data from existing company data, which makes them unsuitable for multi-variant manufacturing.

The project aims to develop a flexible, intelligent assistance system to support small-scale manual assembly processes that can be used effectively in practice. The workers are to be trained and supported in the variant-rich assembly process. The support is provided through the instructions for assembly as well as through warnings in case of an incorrect procedure. In addition, the assistance system can document the assembly sequence.

Reliability, flexibility, cost-effectiveness and feasibility of the assistance system are key factors that are addressed. Furthermore, the assistance system is to be fed from company data and to further enrich this data in automated manner. This is to be achieved by generating, processing and storing the data required for the machine learning of an assistance system without any manual intervention.

The assistance system consists of a number of cameras whose field of view covers elementary areas of the workplace. The recorded data is processed by an industrial PC and fed into the company’s infrastructure. From the infrastructure, this data can be retrieved by other devices, such as augmented reality glasses for displaying occurrences at the workplace.