Improving Special Purpose Machine User-Interfaces by Machine-Learning Algorithms

摘要

This paper proposes to make complex production machines more user-friendly. Improved machines help the operator in case of an error message or a process event by displaying recommendations, such as "in the last 10 occurences of this event the operators performed the following keystrokes". The messages are generated from statistical data on former user interaction and previous process-events. The data represents the knowledge of all the machine operators. The data is gathered by logging user-interaction and process-events during regular operation of the production machine. This approach allows to store the operators' expert knowledge in the production machine without human intervention.