Tsai, Yu-HanYu-HanTsaiLi, Hsu-MingHsu-MingLiCHUN-YEON LIN2025-11-252025-11-252025979833153342721596247https://www.scopus.com/record/display.uri?eid=2-s2.0-105018745166&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/734118This paper presents a permanent magnetic (PM) tracking system utilizing a PM and a Tunnel Magnetoresistance (TMR) sensor array for developing ophthalmic surgical training devices for doctors. The advantage of the PM tracking system is its ability to track an expert surgeon's movements in real time without causing obstructions. The recorded expert trajectories are incorporated into virtual simulators, enhancing the efficiency of surgical training. The Levenberg-Marquardt (LM) algorithm is employed to estimate the location and orientation of the permanent magnet (PM). To predict future PM trajectories, the proposed method compares the similarity between historical and reference trajectories using dynamic time warping (DTW), and this information is further integrated into the Unscented Kalman Filter (UKF) as a dynamic model. The trajectory estimation method, utilizing Dynamic Time Warping (DTW), provides instructional guidance to surgeons during training.false[SDGs]SDG4conference paper10.1109/AIM64088.2025.111757732-s2.0-105018745166