基于神经网络的车辆急加速工况驾驶性评价研究

Research on Driveability Evaluation of Vehicles on Tip-In ConditionBased on Neural Network

针对驾驶性评价中主观评价一致性差及客观评价无法反映人体主观感受的问题,基于对急加速工况的特点分析,构建了急加速工况的驾驶性客观评价体系,并使用BP神经网络搭建驾驶性主观评分预测模型,建立主客观评价间的映射关系。最后,通过实车试验得到驾驶性主客观评价数据集,对神经网络进行训练和测试。结果表明,预测模型的整体准确率在95%以上。

Subjective evaluation in drivability evaluation has shortcomings of poor consistency and the inability of objective evaluation to reflect people’s subjective feel.To solve this problem,the objective evaluation system of tip-in condition is constructed based on the analysis of the characteristics of Tip-in conditions,and the prediction model of subjective score of driving performance is built by using BP neural network,and the mapping relationship between subjective and objective evaluation is established.Finally,the data set of subjective and objective evaluation of driveability is obtained through real vehicle test,and the neural network is trained and tested.The results show that the overall accuracy of the prediction model is more than 95%.