Cognitive Factories: Modeling Situated Entropy in Physical Work Carried Out by Humans and Robots

Entropy in factories is situated. For example, there can be numerous different ways of picking, orientating, and placing physical components during assembly work. Physical components can be redesigned to increase the Information Gain they provide and so reduce situated entropy in assembly work. Also, situated entropy is affected by the extent of knowledge of those doing the work. For example, work can be done by knowledgeable experts or by beginners who lack knowledge about physical components, etc. The number of different ways that work can be done and the knowledge of the worker combine to affect cognitive load. Thus, situated entropy in factories relates to situated cognition within which knowledge is bound to physical contexts and knowing is inseparable from doing. In this paper, six contributions are provided for modelling situated entropy in factories. First, theoretical frameworks are brought together to provide a conceptual framework for modelling. Second, the conceptual framework is related to physical production using practical examples. Third, Information Theory mathematics is applied to the examples and a preliminary methodology in presented for modelling in practice. Fourth, physical artefacts in factory production are reframed as carriers of Information Gain and situated entropy, which may or may not combine as Net Information Gain. Fifth, situated entropy is related to different types of cognitive factories that involve different levels of uncertainty in production operations. Sixth, the need to measure Net Information Gain in the introduction of new technologies for embodied and extended cognition is discussed in relation to a taxonomy for distributed cognition situated in factory production. Overall, modelling of situated entropy is introduced as an opportunity for improving the planning and control of factories that deploy human cognition and cognitive technologies including assembly robotics.