Assessment of Natural User Interactions for Robot-Assisted Interventions

Abstract

Abstract Robotic assistance in clinical interventions provides high precision and performance. However, the acceptance of such systems is still very low. The idea of collaborative robotics promises practical solutions for this problem. To further promote these ideas in the medical domain, novel concepts for user interactions are needed. This work presents a preliminary study on the recognition accuracy of touch gesture interaction with a KUKA LBR iiwa robotic arm. A recognition application utilising a set of 4 different touch gestures was implemented and evaluated by eight participants. The overall recognition accuracy of the system is 89.8%.