A Traffic Safety Early Warning Method Based on Internet of Vehicles

Abstract

AbstractTraffic accidents cause huge casualties and economic losses every year. The main causes of traffic accidents are the untimely response and improper disposal of drivers. The Internet of vehicles (IOT) system is a large system network based on the intranet, Internet and mobile Internet. According to the agreed communication protocol and data exchange standard, it carries out wireless communication and information exchange between vehicles and roads. It provides technical support for the realization of vehicle danger identification and creates new opportunities for safety early warning. Therefore, from the perspective of assistant driving, this paper discusses the dangerous state identification and safety warning methods of vehicles in different traffic scenes in the environment of Internet of vehicles. In this paper, low order car following model is studied for road traffic scene, and response time and following distance characteristic parameters are obtained by using different calculation methods. Through the above parameters, the curve of vehicle speed and following distance can be obtained. This paper analyzes the speed and following distance curves of all vehicles in the fleet, discusses the representativeness of different mean values to the data as a whole, and determines the safety threshold of following distance in normal driving. When the following distance of vehicles is less than the safety threshold, it is judged as a dangerous state; and according to the relationship between the actual value and the safety threshold, it uses the piecewise function to calculate the probability of vehicle collision accident. The algorithm is verified by the traffic data provided by ngsim project, and the results show that the method can accurately describe the dangerous state of vehicles and the possibility of collision accidents. The experimental results show that in the two traffic scenarios of road section and intersection, this paper uses the low-order car following model and the dangerous state recognition method combining time conflict zone with space conflict zone to realize the dangerous state recognition of vehicles, and expresses it by probability. The dispersion coefficient of arithmetic mean is 0.1840. The utility based early warning method can reduce vehicle collision accidents.KeywordsInternet of VehiclesLow order car following modelSpeed and distance curveTraffic safety