Intelligent Assistant System
Artificial Intelligence in System Specification
In product development, it is becoming increasingly difficult to access existing knowledge: the number of projects and stakeholders involved is growing. Additionally, the information about a product along the V-model is constantly increasing. Furthermore, the results themselves are often not consistently documented or considered for future projects. This has significant implications, particularly in the automotive industry, as fixing errors later on can be costly or even impossible. The Fraunhofer IEM, in collaboration with the automaker Nissan, has found a solution to this problem.
By utilizing artificial intelligence techniques, product information can be provided during the specification phase of new products. This is made possible through an intelligent assistant system.
How does the intelligent assistant system work?
During the specification of a new system, elements that have been used in previous projects, such as a camshaft, are employed. The knowledge about these elements exists in unstructured data and serves as the basis for the assistant system. Information from test results and error reports is interpreted using artificial intelligence techniques, such as natural language processing, and translated into a graph database.
Subsequently, the data is interconnected, weighted, analyzed, and made usable for product development. Information about errors that have occurred in the past with system elements is provided as a tool extension that retrieves information from the knowledge base.
The intelligent assistant system checks whether there are any previous experiences or knowledge about the selected component and displays errors or warnings accordingly. In this way, Nissan increases efficiency in product development and supports its employees in their work.