Dynamic System Identification for Robot-Assisted Transcranial Magnetic Stimulation

Robotic-assisted Transcranial magnetic stimulation has been technically developed and widely used to enhance the long-term outcome of brain stimulation. In this paper, the overall dynamic model of the robotic Transcranial stimulation has been successfully analysed; nevertheless, its parameter estimation leads to a significant challenge due to the complicated dynamic nature of the system. To solve this limitation, system identification based on auto-regressive moving average with an exogenous signal model was consequently implemented. The predicted model was carried out and shown to be an effective data matching with the best-fit percentage of 97.10 %. The optimized gains of proportional and integral control were simulatively achieved using Ziegler–Nichols tuning method incorporated with the transfer function of the proposed dynamic model. These algorithm variables were subsequently used in the robotic implicit force control to regulate the contact forces, in which the human’s head is tracked by a 3D scanner in real-time. The results demonstrate significant satisfaction with the overall qualitative performance of the robotic control, and it simplifies effective tracking of the contact force target by indicating closely the desired input with root mean square error of 0.014 N. Hence, it can be concluded that the implicit force control associated with the human’s head detection system, is considered acceptable for the effective Transcranial magnetic stimulation to facilitate safe and reliable physical collaboration between the human and robot.