Real-Time Detection of Erroneous Behavior for a Spindle Exchange Task in IPS2

Abstract

The consideration of human factors for Industrial Product-Service Systems (IPS2) is an essential purpose in the SFB/TR29 to prevent errors due to malpractice or human error. These innovative systems, integrating both products and services in industrial applications, are multifaceted and dynamic, providing a challenge to the operators because of their high complexity. On the operational level this requires human operators to handle highly complex technical systems in different environments. At the same time operator errors pose a considerable risk to the overall system robustness. This negative impact on work complexity, resulting in additional costs, can be counterbalanced by a cognitive system which detects erroneous behavior before damage or injury occurs. This article describes a pretest with this recognition system where a spindle dismantling is performed. The task, system setup and an estimate of the performance for real-time use is reported. © Springer International Publishing Switzerland 2014.