Regression analysis on speed distribution characteristics of ordinary road
Jian Lu,Xianglong Sun,Yue Dai
DOI: https://doi.org/10.3969/j.issn.1001-0505.2012.02.034
2012-01-01
Abstract:The regression models between the average speed, 15th percentile speed, 85th percentile speed and 6 factors, the number of the lanes, the value of the speed limit, the position of the lanes, the vehicle type, the road saturation and the proportion of the large vehicle, are established by the multiple linear regression, respectively. Then, the speed selection model is built by the Logistic regression. The analysis results show that for the two-way six lanes, the minimum speed limit of the passing lane can be controlled according to the 15th percentile speed, and the maximum speed limit can adopt the variable speed limit, the speed limit for different lanes and the speed limit by types of vehicles according to the 85th percentile speed. The large vehicles in the kerb lane often select low speed driving when the saturation is high, while the small vehicles in the two-way two (or four) lanes often select high speed driving when the saturation is low. Different speed limit facilities are set according to the probability of selecting high speed driving. These two regression models can be used to figure out the 15th percentile speed, the 85th percentile speed and the probability of selecting low or high speed driving under specific road conditions, providing a basis in choosing the speed limits and the corresponding facilities for ordinary roads.