Page - 57 - in Joint Austrian Computer Vision and Robotics Workshop 2020

Figure 2. Simulation of the lab use case in ema Work De- signer. boxes, the MB in Table 1 allows for multiple selec- tionswithinspecificmodificationdimensions. In this example, the typeofsafetydevice fordistancemoni- toringallowsforcombiningdifferentdevices forone system variant. For example, one use case variant could be de- fined as follows: The product type A of the rocker lever is assembled in the work system. The collab- orative task of positioning the levers on the trestle is doneinanergonomicheight foranoperator. Trestles aresuppliedbyafeederandmanipulatedbyaUR10. Theinsertion taskof theadjustingbolts iscarriedout by the human while the tightening task is done by the UR3. The velocities of both robots is set to 500 mm/s. The safety function in both areas A and B is basedonforce limitationanddistancemonitoringby software-based workspace limitations. The distance between robot base and operator should be as large aspossible during thecollaboration. 3.1.RepresentedSystemVariants Themaineffectsonpersonalsafetyresultingfrom the selection of system parameters via MB are de- scribed in the following. Impact of Resource Allocation The modification space related to the given resources spans all possi- bilitiesbetweenmanualprocessingtoanalmost fully automated scenario. Special safety considerations arerelevant for thosecaseswherearobot isallocated to a task. For this purpose, all boundary cases must be evaluated separately for e.g. critical contact situ- ations, safety distances as well as force and pressure impacts on the involved body regions of the human. This can lead to restrictionswhichare storedasa set of rules for apartiallyautomatedassessment. Figure3.Setupof theusecase in labenvironment. Impact of Safety Device Several extrinsic safety devices listed in Table 1 are exchangeable, e.g. whether a light curtain or a laser scanner is used for distance monitoring is usually irrelevant. However, the plane used to determine the safety distance has a significant influenceonpermissibledistancesandthe velocity to the moving robot. Horizontal measuring safety devices, such as a safety mat or a laser scan- neronamobilemanipulator,haveasubstantiallydif- ferent information content than vertical devices such as a light curtain. In contrast to horizontal safety devices, vertical safety devices have a higher uncer- tainty in determining the location of humans. How- ever, a safety mat can be partially skipped by a hu- man, whereas a laser scanner mounted on a mobile robot system,canbeused invariable locations. Impact of Workpiece Supply The feeding of the workpieces mainly influences mechanical safety characteristics,whichcanbedeterminedbymeansof a risk assessment. Therefore, the type of the feeding system has no significant effect on the safety related systemvariant. 3.2.SimulatedandExperimentalSetup The virtually designed and simulated laboratory use case is shown in Figure 2, whereas the physi- cal setup of the use case is shown in Figure 3. All required modification dimensions were taken into account, which gives the impression that unneces- sary redundancies exist especially for the listed ex- ternalsafetydevices. However, theseallowustoper- formspecificstudiesonmeaningful combinationsof safety devices and a direct and detailed comparison between them. In order to assess the effects of modifications on 57