A Motion Capture-Based Approach to Human Work Analysis for Industrial Assembly Workstations

In industry, manual work is becoming increasingly important despite high labor costs due to the trend towards smaller productionProduction volumes and a higher number of variants. In order to identify optimization potential and increase productivity, there is a strong need to analyze and understand manual processes. Especially in Small and Medium-sized Enterprises (SME), which have limited resources for classic process time analysis, these are rarely available. In this work, a method for automatic generation of process time analyses of manual assemblyManual assembly processes is presented. It employs an industrial human activity recognitionHuman activity recognition system. Human motion data is captured in the industrial environment for manual assemblyManual assembly operation. It is post processed using a spatial partitioning approach. The study shows about 76% of the manual operations in the proposed use case scenario are automatically detected and the remaining are hardly identified. It is shown that process time analysis can be carried out without expert knowledge and without significant manual effort and provides knowledge about the process, which can be used to identify optimization potentials to increase productivity.