Research Of Rollover Evaluation Index Based On Multi-parameter Vehicle Model With Driving Intention Identification

With the continuous improvement of the automobile industry,the production of automobiles was increasing rapidly.With the increasing popularity of vehicles on the road,some public safety problems such as traffic accidents could no longer be ignored.For SUV and trucks,rollover accidents should also be considered.From analysis of the rollover accidents,the reason for its occurrence was mostly due to the wrong operation of the driver,so the research on the driver behavior characteristics was very urgent.The behavior of a driver depended on his intention,so the prediction of the driver's intention was the most important aspect in the rollover research.This issue mainly focused on the rollover research under the condition of high-speed operation.The main works of this paper were as follows:(1)Set up the dynamic rollover model based on dynamic multi-parameters.Based on vehicle dynamics analysis,a three-DOF vehicle model was set up.Based on the dynamic lateral tire characteristics,a dynamic tire model was set up.Based on the lateral acceleration feedback and the driver's behavior change under different sports conditions,a preview pilot model was set up.(2)Construction of a driver's intention recognition unit.Select the steering wheel angle and the change rate as the characteristic amount of driving intent recognition.Based on the HMM model,the characteristic parameter set in the sample database was selected to identify the driving intention.Then,the parameters were trained and optimized to obtain the result.(3)The GM(1,1)model in gray theory is used to predict the driving behavior.By observing the current observable steering wheel rotation time series,the driving behavior prediction model could be set up to predict the driving behavior.Combined with the recognition result of the driving intention condition in Chapter 3,a new driving behavior prediction model was established.Based on it,improved the driver model in Chapter 2 and made it a preview-predictive driver model.(4)Based on the analysis of rollover mechanism,rollover evaluation index LTR based on vehicle model was deduced,combined with the preview-predict driver model in Chapter 3,the estimated value of rollover evaluation index ELTR was estimated.Based on the active steering control strategy and fuzzy PID controller,a new rollover warning system was designed and verified in high-speed conditions.