Review of machine learning in traffic accident detection

Abstract

With the development and wide application of New Energy Vehicles (NEVs) in the past decade, the chip and power supply method for Machine Learning (ML) intervention provides a good hardware premise, while the 5G technology under the auspices of large-scale rapid data processing is also a prerequisite for the development of Auto Driving, moreover the current traffic accident prediction algorithms have a more critical role in the development of the field of Auto Driving. In this paper, I will consider these aspects as shown below. On the one hand, Anticipation of Traffic Accidents (ATAs) can be well used in todays Autonomous Driving has not yet been popularized, the existing warning information into the corresponding automated driving intervention signals to achieve the possibility of improving the safety of automated driving to increase the possibility of future use of automated driving. On the other hand, the use of ML in the ATAs can be better in the complex traffic environment in a timely manner for the accident warning to achieve a certain degree of reduction in the rate of traffic accidents, in order to optimize the traffic and reduce the economic loss of people. At the same time, in this review, we will propose the combination of data from in-vehicle cameras and road surveillance cameras to analyze the current development of autonomous driving.