Environment sensing for the creation of work cell models

Since small and medium enterprises (SME) mostly produce small lot sizes, industrial robots cannot be applied profitably. This is due to the fact that the efforts for commissioning, such as expert knowledge, set-up time and know-how—compared to manual manufacturing—lead to an unbalanced cost benefit ratio. Hence, commissioning methods need to be developed, providing special commission processes within SMEs. A promising approach is the combination of the individual advantages of the online and offline commissioning in order to support untrained operators. Those hybrid commissioning methods assure the robot motion by simulation, whereby a virtual model of the work cell is required which includes exact solid models for all objects (e.g. robots, machines, fences) in the workspace of the robot. Within SMEs, those solid models are often available for a few objects only. Therefore, the required work cell model cannot be constructed using conventional modeling methods. This paper presents an approach for combining measured 3D depth data with accurate CAD models.