An Adaptive Industrial Human-Machine Interface to Optimise Operators Working Performance

Abstract

Adaptive User Interfaces (AUI) have the potential to deliver advantageous solutions for a wide range of industrial applications. Their ability to adapt to operator interaction patterns and achieve a more personalised interaction can lead to greater efficiency and productivity in the manufacturing process. However, in certain industrial contexts multiple operators interact with the system, rendering it impossible to detect each individual and propose an operator-personalised adaptation. In this paper we propose a data-driven methodology to generate temporal adaptation rules for a multi-operator industrial process. Through the use of machine learning (ML), the methodology: i) analyzes the interaction of different operators with the same machine, ii) selects the most representative adaptations, and iii) generates a set of temporal adaptation rules. The methodology was validated in a real industrial setting resulting in over 40% shorter operator interaction time, and almost 60% number of clicks reduction, thus decreasing the occurrence of interaction errors.