Research on a support-vector-machine-based Variable Speed Limits control model

This paper studied the existing freeway main line variable speed control mechanisms. Since the Variable Speed Limits (VSL) control for freeway main line is a nonlinear system, it is difficult to predict its behavior using quantitative model. Based on the research for Support Vector Machine (SVM), the paper proposed a novel model which used support vector to build segmental variable speed control mechanism in terms of freeway main line segmentation. The definitions of model and training algorithm have been proposed. The simulation results showed that this model´s structure could be determined easily, possessed strong generalized performance. This approach can enhance the controlled area´s traffic flow and improve the traffic situation of freeway main lines.