People with disabilities are still strongly underrepresented in the free labour market. However, having a job is not only essential for self-esteem and one's own financial independence, but also creates personal added value on many social levels. The transition to "Industry 4.0" in particular could make the inclusion of people with mental and/or physical disabilities even more difficult.
Due to the advancing individualisation of production, it is becoming increasingly important to be able to learn new tasks flexibly. What support possibilities collaborative robots - i.e. robots that can work together with humans without a protective fence - offer in this context is currently being investigated by OFFIS scientist Sandra Drolshagen (Production research area) as part of her dissertation.
For this purpose, a robotic assistance system has been developed that uses a depth camera to record the current work steps of the human and gives various hints through the robot as soon as an error has been detected. The hints vary from a simple wave of the robot's arm to various pointing gestures to the robot "demonstrating" the task or calling in a supervisor. Voice instructions were also added to the assistance system in a second step. The assistance system decides for each person individually which assists are carried out at which point in time.
The assistance system was tested and evaluated in two successive studies conducted at the Gemeinnützige Werkstätten Oldenburg. Subsequently, the long-term effects were also investigated. Initial results show that the participants were able to place significantly more components correctly when they were supported by the system. Some of the participants were even able to complete their task correctly only with the support of the assistance system. In addition, various comments directed at the robot indicate that it was perceived as a tutor. For example, it was called a friend or asked what the mistake was. Overall, the cooperation with the robot was perceived as very positive.
Long-term learning effects could also be observed during the second part of the study, in which the participants were accompanied for several weeks. Not only did the people develop their skills further - the AI-supported robot was also able to adapt its assistance to the individual needs of the users.
Contact