Research On Vehicle Driving Training System Based On Virtual Reality Technology

In the era of rapid development of national economy,people use more vehicles,and the traf fic safety problem becomes increasingly serious with the number of vehicles.The increasing demand of automobile makes the traditional driving training industry develop rapidly.However,the weakness of traditional driving training methods are increasingly prominent,such as the difficulty of booking time,the lack of driving vehicles and the short learning time.Meanwhile,the training of students will be affected by external factors such as weather and venue,which greatly reduces the learning efficiency.In addition,this large-scale training method will also consume more fuel,cause serious environmental pollution,and waste a lot of human and material resources.In short,the traditional driving training methods,with low training efficiency,high resource consumption and serious environmental pollution,no longer conform to the contemporary development concept,and it is urgent to find new ways.At present,the non-immersive virtual reality driving training system has been widely used in the field of driving skills training and traffic safety education.This method based on driving training has the advantages of convenient and simple operation,and it can restore the real training site with realistic driving scenes,improve learning efficiency and user experience,and provide new driving training approach.However,the non-immersive virtual reality training system also has obvious shortcomings.First of all,non-immersive system users can’t observe things around them like the real world,and they are easily disturbed by external environmental factors.Secondly,this kind of training system can’t get the coach’s guidance during the training,and it is easy to have improper operation and low learning efficiency.In addition,the lack of coaches to correct students’ bad driving habits in time will inevitably lay a great hidden danger for future road traffic safety.To deal with the abovementioned problems,based on virtual reality technology,we design and develop a new training system for vehicle driving test.Starting from the overall function of the system,based on virtual reality technology,and taking unity 3D as the development platform,this paper developed a vehicle driving simulation training system.We first construct a three-dimensional scene of dynamic road including vehicles and pedestrians,and then combine the force feedback steering wheel components,VIVE helmet,head/hand motion capture,so as to realize the natural interaction between driving action and three-dimensional scene.The construction and rendering of the 3D scene of the system adopts Unity 3D as the software development platform,and combines the driving simulator with the force feedback steering wheel component to collect the steering wheel,gear,brake,throttle,clutch and other data to control the driving state of the vehicle in the virtual scene and drive the camera perspective.We use HTC vive as the display device of HMD system,combined with Leap Motion,an optical based somatosensory interaction device,to capture the user’s hand movements and realize natural human-computer interaction.Finally,in order to better guide students to learn driving skills and traffic safety knowledge without a coach,correct students’ inappropriate operation behavior in time,improve learning efficiency and cultivate good driving habits,this study proposes a novel driving behavior decision-making model based on deep autoencoder network and XGBoost.The model can predict driving behavior through the collected information,and compare the predicted results of the system with the driving behavior of the students.In case of illegal operation,the system and feedback information can help the students adjust driving behavior in time and form good driving habits.Furthermore,this system can also carry out risk analysis on the driving behavior of students.Through the verification and analysis of the system,the driving test training system based on virtual reality proposed in this paper has the following three advantages:1)it can solve the problems in the training process of non-immersive virtual driving test;2)Its driving training assistant system can guide the students to learn driving better;3)It can collect the historical data of driving training to evaluate and analyze the driving ability of students.