Control method of intrusive miniature axial flow blood pump based on deep reinforcement learning algorithm

The invention discloses a deep reinforcement learning algorithm-based intrusive miniature axial flow blood pump control method, which comprises the following steps of: constructing an intelligent agent interaction environment according to a working condition environment of an intrusive miniature axial flow blood pump; determining a state space and an action space of a strategy corresponding to the intelligent agent, and designing a proper reward function; establishing and training a blood pump control strategy based on a deep reinforcement learning algorithm, and solving a blood pump rotating speed optimization problem; and extracting the trained strategy model, and performing network strategy verification according to the difference between different natural heart activity states and aortic pressure values, the hemolysis index and the platelet activation level by means of a blood pump control strategy based on a deep reinforcement learning algorithm to give an optimal solution to the rotating speed of a bloo