一种基于WSN和深度学习的医疗设备定位管理方法

Management Method of Medical Equipment Positioning Based on Wireless Sensor Networks and Deep Learning

目的为实现医疗设备的科学管理,及时对医疗设备进行定位,快速获取医疗设备的位置,本文开发设计了一种医疗器械相对定位方法,实现了对医疗设备的相对定位。方法通过无线传感器网络(Wireless Sensor Networks,WSN)定位技术利用位置已知的少量传感器节点构建WSN来确定网络中目标节点位置信息,采用窄带低功耗的LoRaWAN技术构建WSN,利用基于深度信念网络(Deep Belief Networks,DBN)的医疗设备定位方法,通过多层的神经网络提取特征,最后通过Softmax得到定位结果。结果通过实验,分析了多种其他定位方法的定位结果、4种DBN的定位结果、不同数量的传感器对定位结果影响。本文设计的DBN_10定位方法在训练集上取得了99%的准确率,在测试集上取得了97%的准确率。结论本文提出的设备定位管理方法有利于实现医疗设备的精细化管理,充分发挥医疗设备最大利用价值,助力医院长远健康发展。

Objective To realize the scientific management of medical equipment and quickly obtain the location of medical equipment,this paper developed and designed a relative positioning method for medical devices.Methods Through wireless sensor networks(WSN)positioning technology,a small number of sensors nodes with known locations were used to construct WSN to determine the location information of target nodes in the network.Narrow-band and low-power LoRaWAN technology was used to construct WSN.The positioning method for medical devices based on deep belief networks(DBN)was used to extract features through multi-layer neural network,and then the location result was obtained through Softmax.Results Through experiments,the positioning results of various positioning methods,the positioning results of four kinds of DBN and the influence of different number of sensors on the positioning results were analyzed.The DBN_10 positioning method designed in this paper achieved 99%accuracy in the training set and 97%accuracy in the test set.Conclusion The method of equipment positioning management proposed in this paper is conducive to realizing the fine management of medical equipment,which can give full play to the maximum utilization value of medical equipment and help the healthy development of the hospital.