부산대학교 도서관

The widespread adoption of Virtual Reality (VR)/Augmented Reality (AR) has increased the demand for versatile controllers for VR/AR. Most conventional controllers for VR/AR have issues with a lack of immersive sensation, comfort, and stability. This study proposes a trans-nail PPG (photoplethysmography) controller using fingertip blood volume changes to enable accurate motion prediction. This controller utilizes the fact that the mainly DC component of the PPG signal varies with the state and movement of the hand and fingers. Based on reservoir computing, this controller also learns the PPG signals under various conditions and predicts the state and movement of the hand and fingers. Since this controller is attached to the fingernail, it does not compromise immersion, comfort, or stability. Three PPG sensor circuits in the trans-nail PPG sensor were designed, prototyped, and assembled with an LED on a nail-sized PCB. The output signals of the PPG sensor for various motions such as thumb raising and lowering, wrist raising and lowering, and wrist twisting were used for the learning and predictions of reservoir computing. As a result, the motion predictions with accuracies of over 90% were achieved successfully.