Although robot-assisted surgeries offer various advantages, the discontinuous surgical operation flow resulting from switching the control between the patient-side manipulators and the endoscopic robot arm can be improved to enhance the efficiency further. Therefore, in this study, a head-mounted master interface (HMI) that can be implemented to an existing surgical robot system and allows continuous surgical operation flow using the head motion is proposed. The proposed system includes an HMI, a four degrees of freedom endoscope control system, a simple three-dimensional endoscope, and a da Vinci Research Kit. Eight volunteers performed seven head movements and their data from HMI was collected to perform support vector machine (SVM) classification. Further, ten-fold cross-validation was performed to optimize its parameters. Using the ten-fold cross-validation result, the SVM classifier with the Gaussian kernel (σ = 0.85) was chosen, which had an accuracy of 92.28%. An endoscopic control algorithm was developed using the SVM classification result. A peg transfer task was conducted to check the time-related effect of HMI's usability on the system, and the paired t test result showed that the task completion time was reduced. Further, the time delay of the system was measured to be 0.72 s. Graphical abstract A head-mounted master interface (HMI), which can be implemented to an existing surgical robot system, was developed to allow simultaneous surgical operation flow. The surgeon's head motion is detected through the proposed HMI and classified using a support vector machine to manipulate the endoscopic robotic arm. A classification accuracy of 92.28% was achieved.
Keywords: Endoscopic control interface; Head movement; Machine learning; Minimally invasive surgical procedure; da Vinci Research Kit.