Prediction of Physiological Tremor Based on Deep Learning for Vascular Interventional Surgery Robot

Physiological tremor seriously affects the operation accuracy of the master-slave vascular interventional surgery robot (VISR), which is very necessary to be eliminated. However, there are some issues in the existing methods. For instant, some methods require the prior knowledge of the prediction horizon for accurate estimation tremor signal. Furthermore, these methods assume the process to be nonstationary in the given prediction horizon. Besides, the phase delay of the system has a great influence on the performance of the surgical operating system. Therefore, the effective tremor signal compensation that can be used to generate the reverse motion signal in real time is needed. The paper proposes a multi-step signal prediction method based on LSTM. Combined with the existing method, the deep learning method improves the accuracy of tremor prediction compared with the other prediction method.