Study On Vehicle Warning System On Expressway Curve Based On Vehicle-road Collaboration

With the rapid development of expressway,the driving safety of vehicle on highway has gradually been taken seriously.Because of the complex road alignment,many accidents occurred in curve sections of highway,and the damage of them was more serious than the consequence of accidents occurred in straight lines.The accidents caused by rollover in curve section can result in a 75% mortality rate even when the driver has taken active safety measures.Thus,it is of great significance for vehicle safety warning in curve sections of highway.First of all,reviewing the existing vehicle safety warning systems,a new vehicle safety warning system on highway curve is proposed based on vehicle-road-environment collaboration.The system can warn real-time for different vehicles in different environment,and it is made up of three parts: data acquisition module,calculation module and warning feedback module.Information of vehicle,road,and environment are collected by data acquisition module and transferred to data processing calculation module.Calculation mode uses the received data to calculate the safety degree of target vehicle and conducts the warning judgment according to warning strategy.The warning feedback module feeds back the warning message to the driver.In order to achieve the function of warning for different types of vehicles,a vehicle identification system was developed using Visual Studio 2010 and OpenCV 2.4.9.Background subtraction is a kind of image processing technology,and it was used to extract the side profile of vehicles.After that,Hu moment was used as a characteristic value to recondite the vehicle according to template matching method.Then,rollover and sideslip state were analyzed respectively,on this basis,the calculation method of vehicle safety degree is proposed.The influence factors of vehicle safety such as pavement adhesion coefficient,vehicle track,center of mass were studied to present their system application reference value.After all,the system security warning strategy was put forward which taking layered warning to vehicle.Finally,the system presented in this paper has been tested and validated.192 vehicles were identified through the vehicle identification software system,and the results show that the correct rate of identification system is about 90%.The results of simulation experiments using Carsim and Trucksim were compared with the results calculated by the proposed vehicle safety warning system,it shows that the warning system can effectively realize the function of warning for different types of vehicles in different environment.