危化品车辆装卸载过程识别的Transformer-RNN模型  

Transformer-RNN Model for the Identification of Hazardous Chemical Vehicle Loading and Unloading Process

作　　者：李晓辉 孙子文[1,2] LI Xiaohui;SUN Ziwen(School of Internet of Things,Jiangnan University,Wuxi Jiangsu 214122,China;Engineering Research Center of Internet of Things Technology Applications of Ministry of Education,Wuxi Jiangsu 214122,China)

机构地区：[1]江南大学物联网工程学院,江苏无锡214122 [2]物联网技术应用教育部工程研究中心,江苏无锡214122

出　　处：《传感技术学报》2025年第2期272-278,共7页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(62173160)。

摘　　要：针对危化品运输过程的偷倒、漏倒问题导致的安全事故,构建Transformer-RNN模型,对运输过程中的运行、装载、卸载三种状态进行识别。首先获取通过安装了传感器的车辆传回的速度、载重、原始AD值等实时数据,通过差分提取速度差、载重差、方向差等特征;其次构建融合Transformer和RNN的分类模型,通过Transformer完成对输入的表征学习,RNN进行学习,自注意力机制突出关键特征;最后由全连接网络输出分类结果。实验结果表明,所构建的模型在危化品车运输过程识别中的准确率、查准率、查全率和F1值均优于现有模型。In view of safety accidents caused by theft and leakage during the transportation of hazardous chemicals,Transformer-RNN model is built to identify three states of operation,loading and unloading in the transportation process.Firstly,the original data such as velocity,load and original AD value are obtained from the vehicle installed sensors,and then characteristics of velocity difference,load difference and direction difference are extracted.Secondly,a classification model integrating Transformer and RNN is established.Transformer is used to complete the representation learning of input,RNN is used for learning,and the self-attention mechanism high-lights key features.Finally,the classification results are output by the fully connected network.The experimental results show that the accuracy rate,precision rate,recall rate and F1 value of the model are all superior to the existing models.

关 键 词：车辆装卸载识别 Transformer模型 循环神经网络 自注意力机制 时间序列 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]