基于模糊线性回归的小轿车运行车速区间预测  

Prediction of Driving Speed Interval of Passenger Car Based on Fuzzy Linear Regression

作　　者：解少博[1] 蒋晓君 魏朗[1] 

机构地区：[1]长安大学汽车学院,陕西西安710064 [2]新疆交通科学研究院,新疆乌鲁木齐830000

出　　处：《公路交通科技》2014年第8期126-130,共5页Journal of Highway and Transportation Research and Development

基　　金：国家道路交通安全科技行动计划项目(2009BAG13A07);新疆维吾尔自治区交通运输厅科研项目(2012-20)

摘　　要：将小轿车在公路上的运行车速用三角模糊数来表征。基于二级公路上30个样本路段的平曲线半径、纵坡度等线形数据和实测车速,利用模糊线性回归方法建立了小轿车第85百分位运行车速区间预测模型。通过另外10个样本路段数据对该区间预测模型进行了验证,结果表明:小轿车运行车速的95%置信区间大都处于模糊线性预测区间之内;预测得到的模糊中心值与观测值的相对偏差和模糊度与观测值的比值两种评价指标均在10%以内。同时,将模糊中心值和线性回归预测值进行了比较,结果表明:模糊线性回归模型的平均绝对误差、平均相对误差和最大相对误差三个指标均优于线性回归模型,达到了更高的估计精度。The driving speed of passenger car on highway is described by triangular fuzzy number. Based on the road alignment data of 30 sites on the second grade highway including the horizontal radius and the longitudinal grade together with the observed vehicular speeds on each site, the fuzzy linear regression is implemented to establish the prediction model of the interval of the 85th percentile driving speed. The model is verified with the data of other 10 sites. The result shows that the 95% confidence intervals of the driving speed are mainly within the intervals predicted by the fuzzy linear model, and the performance indices including the relative error between the fuzzy center value and the observation value as well as the proportion between the fuzzy degree and observation data are both within 10%. Additionally, the fuzzy center value and linear regression prediction values are compared, the result shows that the fuzzy linear regression model has a better performance in the mean absolute error, the mean relative error and the maximum relative error, and thus it achieves a higher prediction accuracy.

关 键 词：交通工程 模糊线性回归 区间预测 小轿车 运行车速 

分 类 号：U491.25[交通运输工程—交通运输规划与管理]