Changing lane probability estimating model based on neural network

Changing lane is one of the methods to reach the destination faster and also could bring more highway traffic accidents. This study through the traffic feature recognition, cluster analysis, similarity measurements and estimation, analyzed the vehicle operation parameter before changing lane, proposed a changing lane probability estimating model which combines the SOM (Self-Organization Map) and BP (Back Propagation) artificial neural network and had passed the test of the Vissim micro traffic simulation data. This model contributes to the dynastic analysis and evaluation for changing lanes in the intelligent transportation system, the traffic accidents reduction. So it´s a critical part for establishing the traffic safe system.