Integrating human cognition in cyber-physical systems: A multidimensional fuzzy pattern model with application to thermal spraying

Abstract

The development of next-generation, intelligent manufacturing relies on the realization of human-cyber-physical systems. This key area of transdisciplinary research seeks to interrogate ways of supporting and integrating human cognition and expertise in complex cyber-physical systems, using modelling methods from artificial intelligence. This paper uses the exemplar of wear classification in thermal spraying to outline how relevant cognitive processes can be elicited and combined with technical variables in a singular, efficient and cognitively plausible modelling framework. To this end, eye tracking and high-resolution voltage measurements were performed in a pilot study, and a representative data set was generated for small data problems. Two multidimensional fuzzy pattern classification models were derived. Results show that both human and technical models are significant and complement one another. While the models still require optimizing, they clearly show that the integrative approach is practicable and fruitful. Our findings call for further transdisciplinary research of the development of cognitive assistance as well as predictive maintenance and quality assurance.