Page - 48 - in Proceedings of the OAGM&ARW Joint Workshop - Vision, Automation and Robotics

manually controlled usage. This shift towards the direct manual guidance of the robot was also measurable in two dimensions of UX: Usability and Performance Expectancy. The first was investigated using the System Usability Scale (SUS). The second describes one’s belief that using the system will help him or her to attain gains in job performance, and was measured using two items which were derived from [4]. Table II shows the increase in the dimensions Usability, Learnability and Performance Expectancy. Fig. 7 - Robot B - Tool During expansion stage 1 experiments (with robot A) the user had to use the Touch Panel 95.4% of the time while manual guidance mode was used only 4.6% of the time. This was due to cumbersome navigation in menus and submenus on the robot teach pendant during the parameterization process. As a consequence a more linear programming approach is proposed for expansion stage 2 which led to the integration of the XROB programming system. Fig. 8 - programming activities Fig. 9 - programming time with / without additional programming effort for parameterization of machine vision algorithms Fig. 9 shows the programming time for robot A (T1) and robot B (T2). Total programming time including machine vision increased while training effect and additional input modalities (FT-sensor powered hand guidance,…) yield a net decrease of programming time. Fig. 10 shows that the small acceptance of the manual guidance input modality in robot A can be increased dramatically if the implementation addresses user requirements and wishes. Fig. 10 - preferred input modalities for robot A and B During the video analysis of robot B the expressed emotions and thoughts of the participants were clustered into several main categories. Qualitative feedback mainly focused on ergonomic details, such as the shape of the handholds on the robot, the positions and drag of the buttons/switches, and the fluency of the manual robot guidance. All of the volunteers pointed out that the robot should actively support them during the teaching process. Main feedback clusters TABLE I AVERAGE DURATIONS OF THE TEACHING PROCESS IN STUDY I AND II INCLUDING THE PERCENTAGE OF BOTH CONTROL MODES Duration (m:s) Robot A Robot B Average Total [SD] 6:25 [2:27] 3:36 [1:03] Remote Control [%] 6:25 [100.00] 1:01 [28.43] Physical [%] 0:00 [0.00] 2:35 [71.57] TABLE III USER EXPERIENCE IN STUDY I AND II INCLUDING PERFORMANCE EXPECTANCY (PE), SYSTEM USABILITY SCALE (SUS), AND ITS SUBSCALES USABILITY (SUS-U) AND LEARNABILITY (SUS-L) Duration (m:s) Robot A Robot B Diff. [%] PE [0-5] 2.40 [1.08] 3.40 [0.89] 1.0 [20.0] SUS-U [0-4] 2.00 [0.73] 2.53 [0.27] 0.5 [12.5] SUS-L [0-4] 1.70 [0.76] 2.60 [0.65] 0.9 [22.5] SUS [0-100] 48.50 [13.99] 63.50 [3.79] 15.0 [15.0] 48