FBG touch sensors in a social robot: toward natural behaviors controlled by physical human-robot interactions

Social assistive robots (SARs) have been widely applied to improve users’ quality of life by encouraging continuous and natural human-robot interactions (HRIs). The robot’s ability to effectively perceive and respond to user actions is essential in achieving these interactions. Sensors play a pivotal role in ensuring these behaviors align with the dynamic needs of users, particularly in sensitive contexts like therapeutic settings. This work presents the development and initial validation of a Fiber Bragg Grating (FBG)-based sensor for touch detection in the hand of a SAR, aimed at maintaining HRIs during therapy. The proposed system involves the integration of FBG sensors into four segments of the robot's hand, enabling precise detection of the touch area. A convolutional neural network (CNN) was employed to classify the touch data, demonstrating accurate performance in identifying contact areas across the hand’s surface. The sensor’s robustness and adaptability to the robot’s instrumentation were also evaluated, highlighting its potential to operate reliably in therapeutic environments. This generates future research into interaction modalities, such as multi-point contact detection and adaptive interaction responses sensed with optical fiber-based technology.