ProdBot - intelligent assistance in production
Motivation
Due to the increasing complexity of the production environment, manufacturing companies face the challenge of providing their operational employees with necessary process information. The research project ProdBot aims to investigate and test the use of chatbots in a production environment. The chatbot is intended to support operational employees in obtaining information in the industrial production environment and thus make decision-making processes faster, more efficient, and more effective.
Objective
The release of the ChatGPT software (OpenAI) illustrates the incredible power of chatbots in private environments. The research project funded by the BFS investigates the use of chatbots in production. The chatbot supports production planners in the operational environment, e.g., in querying expected delivery dates or the availability of preliminary products.
Research Procedure
The project will develop a methodology for the use of chatbots in production.
The concept addresses the existing challenges in using digital chatbots in the production environment with new methods for knowledge storage, language understanding, and data integration. Examples are the support of employees in order control by a fast and natural language-driven analysis of the current workload and order situation or in maintenance by an efficient provision of existing knowledge.
Furthermore, new methods are to be developed that enable the chatbot to learn intelligently based on user interactions, the ontology to be developed, and new data. An ontology is a "map" that organizes and defines the concepts and their relationships within a specific domain (e.g., production). The learning capability refers, on the one hand, to improving language understanding and user interaction and, on the other hand, to the intelligent extension of the ontology. Based on the methodology, demonstrators are created in two different industrial use cases for the practical implementation of the chatbot.