Intersection vehicle collision early warning method based on Internet of Vehicles environment system

Abstract

The invention provides an intersection vehicle collision early warning method based on an Internet of Vehicles environment system. A remote service center firstly judges whether a first vehicle and a second vehicle are both in the same intersection range and both drive to the intersection at the current moment; if yes, the remote service center carries out prediction through the optimized first vehicle LSTM neural network to obtain a first vehicle prediction batch sample at the current moment, carries out prediction through the optimized second vehicle LSTM neural network to obtain a second vehicle prediction batch sample at the current moment, and carries out calculation according to a collision risk calculation model to obtain a collision risk time sequence; and then each sequence value of the collision risk time sequence is compared with a risk threshold value, and if the collision risk is greater than the risk threshold value, the remote service center controls the first vehicle and the second vehicle to slow down to avoid the collision risk. Compared with the prior art, the method has the advantages of long vehicle driving state prediction time delay, no need of manually setting vehicle driving parameters, scientific risk assessment mode and high early warning precision.

Description

technical field [0001] The invention belongs to the technical field of Internet of Vehicles and deep learning, and in particular relates to a vehicle collision early warning method at an intersection based on an Internet of Vehicles environment system. Background technique [0002] With the rapid development of economy and society and the improvement of people's living standards, the number of motor vehicles in my country is increasing rapidly. According to the latest data, as of the end of 2018, the number of motor vehicles in my country has reached 327 million, and the number of motor vehicle drivers has exceeded 400 million for the first time, reaching 409 million. With the rapid growth of the number of motor vehicles, the occurrence of traffic accidents is also becoming more frequent. In 2018 alone, there were 245,000 traffic accidents in my country, 63,194 people died, and the direct property loss was 1.38 billion yuan. Among them, traffic accidents occurring at inter...

AI Technical Summary

Problems solved by technology

As a typical representative of recurrent neural network, the standard RNN neural network can effectively deal with nonlinear timing problems, but there are still problems of gradient disappearance and gradient explosion, which make it impossible to deal with timing problems with too long delay.

Images