Development of User Profiles for Adaptive Assistance Robots

Abstract

Due to the aging population, with an accompanying shortage of nursing staff, the idea of an independent life as long as possible with the help of technical aids is becoming more and more attractive. One way to achieve this is to use autonomous robots that support seniors in their everyday lives. The professional use of such an assistant robot in care facilities to support the care staff would also be conceivable. For this purpose, this thesis deals with the creation of user profiles for individual communication and interaction of assistant robots with users of the target group seniors. Seniors are a very diverse target group because with age many impairments, such as hearing, visual, motor skills and speech impairments can be associated. For this purpose, a concept for creating user profiles was developed. When a user first encounters the assistance robot, the robot collects the user characteristics and saves them in a user profile created especially for the user. This feature collection is done implicitly by the robot and explicitly by means of a question-answer dialogue between robot and human. Since it is not yet known at the time the characteristics are ascertained which modalities the assistance robot uses to best interact with its counterpart, this is done via as many channels as possible. Once the user profile is complete, it is saved as a JSON file. The result of the development of user profiles and the possibility of using them to implement an adaptive behavior of assistant robots is the increase in usability by promoting the accessibility of the entire system to its users.

Download Thesis

You can download the full submission as PDF file (6,3 MB) here.