Vascular interventional surgical robots (VISRs) can help doctors to avoid X-ray radiation. This paper proposes a leader–follower isomorphic robot where the structural form and operational logic are completely identical. The doctor’s operation on the leader robot is precisely replicated on the follower robot, enabling delivery and rotation capabilities. It can further achieve collaborative operation. This control system adopts a four-channel scheme based on acceleration and can achieve approximately ideal transparency. The leader–follower delivery error of the catheter/guidewire is less than 1 mm, and the leader–follower rotation error of the guidewire is less than 0.3° in an actual intervention task based on a human vascular model. Subsequently, the cumulative sum (CUSUM) method was used to evaluate the learning curve of the robot system, demonstrating that both operators could master the operation method within 10 trials. We classified operators with different operational experience using machine learning methods. The classification process includes time-frequency domain feature extraction, feature selection based on the Relief method and random forest (RF) method, and a BP neural network (NN) classifier. The results indicate that this method can achieve accuracy of 94%. This paper comprehensively discusses the robot system from the perspectives of the mechanism design, control methods, and evaluation methods, providing valuable insights for the design of related robotic systems.
Read full abstract