Y.C. Ding, W.J. E, Y.M. Liu, S.F. Zheng, S.C. Xu

Pages: 15-30

Abstract
The collaborative lane change control effect of multi lane vehicles is of great significance for the safety protection of vehicle driving. In order to reduce the collision frequency of multi lane vehicle collaborative lane changing, a multi lane vehicle collaborative lane changing control method considering driving characteristics is proposed. This method utilizes a driving characteristic classification method based on the Newell model and BP neural network. After extracting the driving characteristic parameters of multi lane vehicles through the Newell model, backpropagation training is used to minimize classification errors. The interlayer connection threshold and bias of the BP neural network are set, and the extracted driving characteristic parameters of multi lane vehicles are input into a trained BP neural network model to classify and recognize different driving characteristics. Combining the characteristics of non driving, a multi lane collaborative lane change method based on driving characteristics is used to construct a multi lane model, analyze the status of one's own lane and target lane, and use satisfaction as the lane change indicator to control the collaborative lane change of multi lane vehicles.The experimental results indicate that the multi lane vehicle collaborative lane change method can effectively overcome the problem of lane change collisions and has high application value.
Keywords: consider driving characteristics; multiple lanes; vehicle collaboration; lane change control; newell model; bp neural network