A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Car Following Prediction Based on Support Vector Regression and Multi-adaptive Regression Spline by Considering Instantaneous Reaction Time
Abstract Car following modeling is one of the major issues in microscopic traffic simulation, and the accuracy and reliability of microscopic simulation models highly depend on the models. This research used a highly efficient method of regression for car following modeling such as support vector regression and compared it with multi-adaptive regression spline method as a nonparametric method. Meanwhile, the factors of “speed” and “distance to the following car” were used as the model’s inputs and it was proved that the accuracy of support vector regression model exceeded that of multi-adaptive regression spline. Driver’s reaction time is another unavoidable factor in car following modeling, which varies based on driver–vehicle features and traffic conditions. The research determines this value with respect to the time lag between the diagrams of the following and leading cars distance, the speed of the following car, and the execution of support vector regression model. The model was then implemented using the microdata of a highway traffic flow in the USA for 3 lanes. The research shows that the proposed model has a proper validity after entering driver’s instantaneous reaction time, which is based on the proximity of the results to the real condition of car following in simulation. A comparison was made between the simulations carried out using traditional models and the models proposed in this research. The results can be considered as a basis for other studies on car following modeling and microscopic traffic simulation in highway field.
Car Following Prediction Based on Support Vector Regression and Multi-adaptive Regression Spline by Considering Instantaneous Reaction Time
Abstract Car following modeling is one of the major issues in microscopic traffic simulation, and the accuracy and reliability of microscopic simulation models highly depend on the models. This research used a highly efficient method of regression for car following modeling such as support vector regression and compared it with multi-adaptive regression spline method as a nonparametric method. Meanwhile, the factors of “speed” and “distance to the following car” were used as the model’s inputs and it was proved that the accuracy of support vector regression model exceeded that of multi-adaptive regression spline. Driver’s reaction time is another unavoidable factor in car following modeling, which varies based on driver–vehicle features and traffic conditions. The research determines this value with respect to the time lag between the diagrams of the following and leading cars distance, the speed of the following car, and the execution of support vector regression model. The model was then implemented using the microdata of a highway traffic flow in the USA for 3 lanes. The research shows that the proposed model has a proper validity after entering driver’s instantaneous reaction time, which is based on the proximity of the results to the real condition of car following in simulation. A comparison was made between the simulations carried out using traditional models and the models proposed in this research. The results can be considered as a basis for other studies on car following modeling and microscopic traffic simulation in highway field.
Car Following Prediction Based on Support Vector Regression and Multi-adaptive Regression Spline by Considering Instantaneous Reaction Time
Pahlavani, Parham (author) / Poor Arab Moghadam, Mohsen (author) / Bigdeli, Behnaz (author)
2018-09-28
13 pages
Article (Journal)
Electronic Resource
English
Travel Time Prediction with Support Vector Regression
| British Library Conference Proceedings | 2003
Travel time prediction with support vector regression
| IEEE | 2003
Urban Arterial Travel Time Prediction Using Support Vector Regression
| Springer Verlag | 2018
Railway Freight Volume Prediction Based on Support Vector Regression (SVR)
| British Library Conference Proceedings | 2014