Touch-Sensitive Hand Interactions for Social Robots Using Fiber Bragg Grating Sensors

Abstract

Physical human-robot interaction (pHRI) has been demonstrated to be essential in the implementation of social assistive robots (SARs), which require advanced sensing capabilities for accurate and responsive engagement. This study presents the development and validation of a fiber Bragg grating (FBG) sensor network integrated into the hand of the CASTOR robot to classify complex pHRIs. Nine pHRIs were collected and evaluated within the high five, pets, handshakes, hits, and pinches categories. Four machine learning (ML) algorithms were tested, and the Bagged Decision Tree Classifier (BDTC) achieved the best performance. During testing, the model achieved an accuracy of 98%. The results demonstrate that the proposed FBG sensor network can classify complex pHRIs. Future work will explore additional instrumented areas of the robot and expand the physical interaction analysis to enhance social robot adaptability and user experience.