基于神经网络的人工心脏温度预测

Working Temperature Predication of Artificial Heart Based on Neural Network

为了实现人工心脏泵的无传感器温度预测方法,该文研究了应用BP神经网络和遗传算法预测其温度的方法。针对人工心脏泵在植入人体后所受到的环境限制,研究通过体外较易测量的参数预测泵体运行温度。对比了BP神经网络的预测精度与遗传算法优化后的BP网络预测精度。经实验验证,出现误差大于1%的概率为1.84%。

The purpose of this paper is to achieve a measurement of temperature prediction for artificial heart without sensor, for which the research briefly describes the application of back propagation neural network as wel as the optimized, by genetic algorithm, BP network. Owing to the limit of environment after the artificial heart implanted, detectable parameters out of body are taken advantage of to predict the working temperature of the pump. Lastly, contrast is made to demonstrate the prediction result between BP neural network and genetical y optimized BP network, by which indicates that the probability is 1.84%with the margin of error more than 1%.