基于电流参数的导管泵的智能生理控制

Intelligent Physiological Control of a Catheter Pump Based on Current Parameters

摘要: 经过半个多世纪发展，叶轮式人工心脏血泵在血液相容性、植入性和耐久性方面已逐步满足临床需求，但患者术后舒适性及血泵流量对扬程变化的响应仍需改善。针对左心室辅助泵难以自动匹配右心室流量的问题，本文提出基于神经网络自动检测血泵参数的仿生控制系统。系统根据电机功率和转速实时推算扬程，基于扬程、流量与电压关系制定补偿控制策略，并采用ML4425驱动芯片与STM32控制器完成硬件设计。结果表明，与等电压或等转速控制相比，该系统提高了流量对扬程变化的响应灵敏度，为维持左右心室流量平衡提供了参考。

Abstract: After more than half a century of development, centrifugal artificial-heart blood pumps have gradually met clinical requirements for hemocompatibility, implantability, and durability. However, postoperative comfort and the sensitivity of pump flow to head changes remain challenges. To address the difficulty of left ventricular assist pumps in automatically matching right ventricular flow this study proposes a biomimetic control system that detects pump parameters with a neural network. The system estimates head from motor power and speed, develops a voltage-compensation strategy from the head-flow-voltage relationship, and implements the hardware with an ML4425 driver and STM32 controller. Results showed higher flow sensitivity to head changes than equal-voltage or constant-speed control, providing a reference for maintaining left-right ventricular flow balance.