针对车辆在运行过程中左右侧车轮垂直载荷难以直接测量的问题，结合卷积神经网络(convolutional neu-ral network，CNN)和长短时记忆神经网络(long short-term memory，LSTM)对横向载荷转移率(lateral load transfer rate，LTR)进行预测。建立重型载货车TruckSim动力学模型，在鱼钩工况、J-Turn工况和双移线工况下采集车辆的侧向加速度、横摆角速度等行驶状态参数。在MATLAB中建立CNN-LSTM模型，利用CNN-LSTM模型的特征提取和时间序列预测功能，对重型载货车LTR值进行预测，并在多种工况下验证CNN-LSTM模型的性能。结果表明:在不同行驶工况、车辆参数及路面条件下，CNN-LSTM模型能够对重型载货车LTR值进行有效预测。
Rollover Prediction of Heavy-duty Trucks Based on CNN-LSTM
To address the difficulty in directly measuring the vertical load of the left and right wheels during vehicle operation,a convolutional neural network(CNN)and a long short-term memory neural network(LSTM)were used to predict the lateral load transfer rate(LTR).By establishing a TruckSim dy-namic model for heavy-duty trucks,state parameters of the vehicle under fishhook,J-Turn,and double lane conditions were collected,such as lateral acceleration and yaw velocity.A CNN-LSTM model was established in MATLAB,and the feature extraction and time series prediction functions of the model were used to predict the LTR of heavy-duty trucks.In addition,the performance of the CNN-LSTM model was verified under various working conditions.The experimental results show that the CNN-LSTM model can effectively predict the LTR of heavy-duty trucks under different driving conditions,vehicle parameters,and road conditions.
heavy-duty trucksrollover predictionLTRtime series prediction
万方数据
维普
