Measuring Interaction Using Levels of Automation over TIme

Predictability of product quality and cycle and lead-times is vital for companies in order to improve production system performance and competitiveness. Ways of increasing predictability are e.g. elimination of identified problems and standardization of tasks. A frequent cause of problems in production systems is humanautomation interaction, a topic frequently approached by several research domains. This paper proposes that by analyzing changes in the production system’s physical and cognitive levels of automation (LoA) over time, system performance as well as effects of improvements can be predicted. A case study was conducted to test a proposed LoA-Time method, where a taxonomy for physical and cognitive LoA assessment was used. Preliminary results indicate that LoA-Time can be used to identify differences in skill and interaction patterns in order to suggest system improvements, in terms of reduced sub-task time and better support.