@inproceedings{kowalski2019a, Author = {Kowalski, Christian and Arizpe-Gomez, Pedro and Weiß, Sebastian and Gliesche, Pascal and Hein, Andreas}, Title = {Intuitive human-robot interaction for physical support during nursing activities using myoelectric signals}, Year = {2019}, Pages = {92}, Booktitle = {Zukunft der Pflege Tagungsband der 2. Clusterkonferenz}, type = {inproceedings}, Abstract = {Nurses are exposed to immense physical strain, which often leads to early retirement from work. Due to the prevailing care crisis, it is therefore important to provide support in this domain to counteract the loss of nursing staff. A large potential resides in the use of robotics for physical relief by supporting during patient positioning activities. In any form of human-robot interaction with assistance systems, the difficulty lies in the ideal selection of concepts for initiating actions on the part of the robot – especially in safety critical environments. For this reason, the present scientific work deals with the implementation of an intuitive interaction concept for non-contact communication with a robotic manipulator attached to the patient bed, which can support both patient and nurse. For such an interaction concept, the robot needs to gather information about when and where to move to in order to support properly. In our case, the communication of when to move is done by recognizing a specific gesture based on myoelectric signals of the nurse's lower arm using a Myo wristband. Moreover, the target location is the nurse's current hand position calculated in three-dimensional space by using image data from external cameras. During nursing activities, the nurse initiates the movement of the robot arm by using a gesture like fin-ger spreading. The results show that this kind of interaction is feasible for positioning movements which require little to moderate force – high-force movements are less likely to be detected reliably with the Myo wristband.} } @COMMENT{Bibtex file generated on }