A Fuzzy-Logic-Based Method for Autonomous Guidewire Delivery

Guidewire delivery is a critical step in percutaneous coronary intervention (PCI). For vascular robotic systems that assist clinical surgery, the development of autonomous guidewire delivery methods can reduce the learning curve and mental burden on doctors, as well as decrease surgical response time. However, current mainstream reinforcement learning (RL)-based methods suffer from a lack of interpretability and generalization, in addition to requiring long training periods. To address these issues, this paper proposes a novel method that constructs a fuzzy logic system based on the thought process of human doctors during clinical surgery when delivering the guidewire. The system calculates the rotational control parameters of the guidewire according to the relative angle of its distal end within the vascular system and historical attempt data, thereby enhancing the guidewire’s capability to navigate complex vascular branches. Experimental results demonstrate that the proposed method significantly improves the success rate and delivery efficiency in target point delivery tasks.