Research On Distracted Driving Recognition Of Truck Drivers Based On Driving Simulated Experiments

Road traffic accidents have always been one of the major threats to the safety of national life and property,large trucks have been ranked in the forefront of many road killers.Driver distracted driving is an important cause of road traffic accidents.It is necessary to conduct in-depth research on it,discuss the influence of distraction behavior on driving operation behavior and vehicle operation,and provide theoretical support for developing truck driver distraction monitoring system,which is of great significance for improving the traffic environment and improving the safety of the human-vehicle-road system.Firstly,based on the investigation of the distraction behavior of truck drivers,this paper obtains the typical secondary task of the truck driver and the scene that is prone to distraction behavior.Then,based on the truck simulation driving experiment platform,the drinking water and voice message chat simulation driving experiment scheme was designed and experimented.Finally,the vehicle running status information of the truck driver during distraction driving is collected and analyzed,and the distraction driving discriminating model of the truck driver is established.This paper provides a theoretical basis for real-time monitoring of truck drivers' driving status and traffic safety accident identification and analysis,and provides some academic value for studying the distracting driving effects caused by different driving secondary tasks.The main results obtained during the study period are as follows:(1)Conduct a survey of truck drivers' distracted driving behaviors.Using natural driving observation(NDS)to video surveillance and video surveillance of nearly 100 freight car drivers in three freight companies in Beijing and Ningbo,and in combination with the driver distraction driving questionnaire,the driver's distracted driving behavior is studied.(2)For the selected typical truck driver distracted driving behavior based on the driving simulation experiment platform,the drinking water and voice message chat distracting driving simulation experiment was carried out.(3)The statistical analysis of the impact of two typical secondary tasks of truck driver drinking water and WeChat voice message chat on driving performance was carried out.The effects of different water cup positions,different drinking time and different complex difficulty levels on the vehicle's motion state characteristics are obtained.(4)The ReliefF algorithm is used to select the seven most important characteristic indicators from the 18 driving behavior characteristics as the discriminating index of the truck driver's distraction state.Then these characteristics are used as the input of the random forest combination model to establishing a discriminating model for the driver's distraction state based on random forest combination algorithm,The discriminant result shows that the accuracy of the model established in this paper is 91.96%.Then,compared with other machine learning algorithms such as decision tree C4.5,AdaBoost-BP and BP neural network,the performance of the driver's distraction state discrimination model based on random forest combination algorithm is better,which can effectively carry out trucks.