Usability of human-robot interaction within textile production: Insights into the acceptance of different collaboration types

Abstract

With the advent of digitization and automation in industry, the use of technological innovations is increasing especially in areas that were previously exclusively processed manually. The interdisciplinary research cluster “Internet of Production” (IoP) has made this very fact its research task precisely to reshape production through technologies such as machine learning or cloud computing. In this way, fully networked processes as well as digital shadows of real production chains and products are expected to achieve major advantages for production companies. However, the reality in most production processes using limp materials (e.g. textiles or foils) is different as these processes are difficult to automate. In textile composite production, for example, about 40 % of all components are manufactured manually. Human experts are skilled in dealing with the brittle textiles and have an excellent sense of materials, enabling them to flexibly manufacture even very complex components. Nevertheless, these manual processes cause problems like slow production speed as well as back or wrist problems for employees due to working in bent positions and applying a lot of force. In addition, the shortage of skilled workers and ageing workforces are compelling companies to react and prevent a loss of existing tacit expert knowledge. Fully automated processes cannot be used for many composite components due to unproportioned costs, limited flexibility or a too high level of complexity. In order to still save time and costs, besides improving working conditions and securing tacit knowledge, the approach of human-robot interaction is promising. A successful and appropriate integration can be achieved by considering the technical possibilities as well as the workers specific needs. The current empirical state of the art on human-robot collaboration mainly focuses on performance, safety and task assignment. So far the human factor remains underexposed. This study therefore focuses on the user and investigates motivational and acceptance-relevant factors that enable a qualitatively valuable division of labour in relation within the human-robot interaction.In a user-centered design the acceptance of the interaction with a cobot during a textile composite production process is examined. The within-subjects experimental design requests participants (n = 21) to form a textile into 3D-shape to the best of their ability. The cobot acts as a technical assistant and supports the execution of tasks in three different levels of autonomy (low, middle, high). Based on the technology acceptance model (TAM), the general acceptance of the interaction is assessed. Furthermore, perceived autonomy, perceived control, bodily security, perceived mental and physical effort is surveyed after each task. It turns out that the intention to use a cobot is generally high. The greatest influence on the acceptance is characterized by hedonic motivation, followed by ease of use, trust and usability. The evaluation regarding perceived autonomy and control over task performance and physical safety increases with decreasing cobot activity. Based on the results, we derive recommendations for an optimal use of a human-robot interaction, in which the cobot is perceived as a supporting component and at the same time the maximum autonomy of the worker can be maintained.