摘要：AbstractManual manufacturing and assembly systems radically changed after the advent of the fourth industrial revolution (Industry 4.0). The production paradigm shift to mass personalization involves the customers since the product design phase. The remarkable variety and complexity of these processes is tackled by highly skilled human operator, which perform value added and non-repetitive tasks. Thus, the virtualization of manual operations represents a great opportunity to monitor, analyze and successively optimize these production processes. Within this context, this research proposes an original hardware/software architecture called Motion Analysis System (MAS) developed to monitor and to evaluate manual manufacturing and assembly processes through motion capture technologies. The MAS adopts a hardware architecture represented by a network of depth cameras, e.g. a marker-less optical motion capture, to digitalize the operator movements and postures at 30 Hz while he performs production activities. A customized software architecture exploits these data in relation to the workstation 3D layout to automatically and quantitatively evaluate a set of productive KPIs. The MAS represents a reliable, powerful and meaningful tool for production managers and practitioners able to provide an in-depth time and space analysis of the activities performed by an operator, as the walking paths, the accessed picking locations, the used space within the workstation and the hand movements. Finally, this contribution ends with a real industrial application of the MAS which analyzes the manual assembly processes of a microwave oven. The system setup is presented and discussed along with the case study key results and findings.
