Asynchronous Robotic Arm System Based on Augmented Reality and SSVEP-based BCI

Abstract

With the development of robot technology, more and more disabled people can improve their external environment's interactive ability. But for some people with severe interaction barriers, the current interaction methods are still difficult to meet their needs. In this study, to improve the situation, an asynchronous robotic arm control system based on steady-state visual evoked potential (SSVEP) non-invasive BCI is implemented using augmented reality (AR) glasses. With simple guidance, the subjects can master the asynchronous control method constructed flexibly by using the pseudo key. An SSVEP based AR-BCI adopts filter bank canonical correlation analysis (FBCCA) with 12 targets, which can be used for subjects without system calibration. The results of online test on 4 healthy subjects showed that the subjects could grasp and place the ball flexibly and freely by choosing two-dimensional coordinates. The time of choosing a single coordinate was 4 s (3 s visual stimulation, 1 s feedback), the classification accuracy rate was 97.17%, and the averaged ITR was 48.23 bits/min. These results indicate the feasibility of asynchronous AR-BCI in robotic arm control. The system also has application and development prospects to provide for the aged and help the disabled.