基于卡车行驶数据的驾驶行为分析及评价模型研究

Research on driving behavior analysis and evaluation model based on truck driving data

采集了重型卡车6万km的运行数据,提出了一种基于机理分析和数据特征融合的驾驶经济性和安全性的评价方法。首先对实际运行数据进行处理,通过控制变量选择6次外界环境相同的车辆驾驶过程,通过机理模型分析提取了3个与瞬时油耗密切关联的特征参数,将其聚类成3个油耗等级,然后建立了一个以瞬时油耗预测为主要目标的经济性评估模型,经过训练、测试后得到该模型的预测准确率达96.3%。从纵向控制角度分析,计算得出3个车辆驾驶安全性指标,对各个车辆的驾驶安全等级进行分类,利用多维度的层次分析法计算得出各自的权重,从而针对车辆驾驶行为进行安全性评价,最后对其所评价的模型结果进行案例验证。

Running data of heavy trucks of 60000 km are collected,and a driving economy and safety evaluation method based on mechanism analysis and data feature fusion is proposed.First,through the processing of actual operating data,six driving processes of the vehicle with the same external environment are selected through the control variables,and three characteristic parameters closely related to the instantaneous fuel consumption are extracted through the mechanism model analysis,which are clustered into three fuel consumption levels.Then,an economic evaluation model with instantaneous fuel consumption prediction as the main objective is established.After training and testing,the prediction accuracy of the model reaches 96.3%.From the perspective of longitudinal control analysis,the three vehicle driving safety indicators are calculated,and the driving safety level of each vehicle is classified.The multi-dimensional hierarchical analysis method is used to calculate their respective weights,so as to evaluate the safety of vehicle driving behavior.Finally,the results of the evaluated model are verified by a case.