Driving-behavior Prediction And Driving-mode Identification Of Pure Electric Vehicle

With the environment problem being more and more serious,more and more people pay attention to the new energy resources and related industry.Meanwhile,manless driving,aided driving,and ITS(Intelligent Transportation System)is developing rapidly in recent years.Thesis identified the driving intention,driving style,driving behavior and driving mode in pure electric vehicle by analyzing the new energy electric bus running data.Driving intention means a certain state the driver wants the car to be in the process of driving.Driving style is same with driving habit.Driving behavior means an operation the driver makes.And driving mode means the driving style to a certain driving intention the driver behaves in certain running condition.Thesis also described the related problem appeared in the process of research.The research content mainly includes the following three aspects:(1)Proposed the two-step filling method to solve the filling problem in the whole record missing situation from the car running data which exists temporal characteristics.The filling accuracy was improved as much as possible and the filling data was in accordance with the automobile driving principle and the fact.(2)Predicted the driving behavior with the accuracy over 90% using MLPNN(multi-layer perceptron neural networks)based on driving intention and driving style;Identified the driving mode using K-Means clustering algorithm based on the car running condition,which conducted good result.Above two works can solve the match problem about the self-driving model that occurs in intelligent driving may be not a match for driver mode.Meanwhile,thesis proposed two evaluation parameters for driving mode to decide the influence from different driving modes and application in parameters.(3)The parameters for identifying the car running condition used vehicle control parameters and running states' parameters,so as to eliminate the limitations of the running condition identified by speed only.The significance of the study is to predict Driving Behavior and identify Driving Mode,and then promote the development of automatic driving/auxiliary driving,so as to improve the Driver Comfort and driving efficiency.Based on the research content,we can conduct many related studies like driving control,driving pattern design,optimizing the driving energy consumption and so on.