手术室护理设备维修需求自动化感知方法  

Automatic perception method of maintenance needs of operating room nursing equipment

作　　者：付佳[1] 田甜[1] FU Jia;TIAN Tian(China Medical University Shengjing Hospital,Shenyang 110000,China)

机构地区：[1]中国医科大学附属盛京医院,沈阳110000

出　　处：《自动化与仪器仪表》2022年第11期286-290,共5页Automation & Instrumentation

摘　　要：针对传统手术室护理设备故障感知效果差、主动预警和自动化水平低的问题,提出基于一种深度Q网络模型的手术室护理设备维修需求自动化感知方法。首先,在深度Q网络中输入基础护理设备数据后,通过该网络对护理设备多维数据进行特征提取和训练;然后利用强化学习对样本故障特征进行自主决策和故障分类结果输出,通过故障主动预警机制进行设备自检和维修。实验结果表明,提出的自主感知模型的感知结果与实际故障感知模型十分接近,两者最大误差仅为0.26,且本模型的自主感知和主动预测精度分别为90.47%和94.85%,均高于传统运维管理模型。由此说明,本系统能够实现护理设备的自动化感知和自主预警,故障感知效果显著提升,自动化水平进一步提高,可在手术室护理领域进行广泛应用和推广。In view of the problems of poor fault perception effect and low active early warning and automation level of traditional operating room nursing equipment, an automatic sensing method of operating room nursing equipment maintenance needs based on a deep Q network model is proposed. Firstly, after inputting basic nursing equipment data into deep Q network, the multidimensional data of nursing equipment is feature extracted and trained;then reinforcement learning is used to make independent decisions and output fault classification results on sample fault features, and conduct equipment self-inspection and maintenance through active fault warning mechanism. The experimental results show that the perception results of the proposed autonomous perception model are very close to the actual fault perception model, the maximum error of which is only 0.26, and the accuracy of autonomous perception and active prediction of this model is 90.47% and 94.85%, respectively, which are higher than the traditional operation and maintenance management model. Therefore, it shows that the system can realize the automatic perception and independent early warning of nursing equipment, the fault perception effect is significantly improved, the automation level is further improved, and can be widely used and promoted in the field of operating room nursing.

关 键 词：护理设备 深度Q网络 维修需求 自动化感知 主动预警机制 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]