Vehicle and pedestrian collision early warning method and system based on Internet of Vehicles

Abstract

The invention relates to a vehicle and pedestrian collision early warning method and system based on the Internet of Vehicles, and the method comprises the steps: converting a speed factor, a vehicle length factor and a human body reaction time factor which affect the collision into a probability model, and obtaining corresponding normal probability distribution; the probability that the person collides at the possible collision point and the probability that the vehicle collides at the possible collision point are obtained, the product of the two probabilities serves as the final collision probability, the final collision probability is compared with the dynamic threshold value, and if the final collision probability is larger than the dynamic threshold value, vehicle deceleration, whistling or flashing are conducted to prompt pedestrians. According to the method, the factors influencing collision are comprehensively analyzed to predict the probability, the method is accurate, the error-tolerant rate is high, direct collision between people and vehicles can be effectively predicted, the purpose of reducing traffic accidents is achieved, and the method can be widely applied to the field of automobile intelligence.

Description

technical field [0001] The present application relates to the field of artificial intelligence, in particular to a vehicle-pedestrian collision warning method based on Internet of Vehicles. Background technique [0002] On urban roads, especially at intersections, pedestrians / non-motorized vehicles run red lights, resulting in collisions and traffic accidents. The current conventional methods for predicting collisions between people and vehicles are as follows: [0003] (1). Vehicles, pedestrians / non-motorized vehicles are regarded as uniform speed / uniform acceleration for collision estimation. The disadvantage is that the speed of vehicles, pedestrians / non-motorized vehicles may change continuously during travel, so that the possibility of collision is also constant. changes, leading to inaccurate predictions. [0004] (2). Describing the collision position of vehicles, pedestrians / non-motor vehicles as colliding at a fixed point, the disadvantage is that relying on the co...

AI Technical Summary

Problems solved by technology

[0003] (1). Vehicles, pedestrians / non-motorized vehicles are regarded as uniform speed / uniform acceleration for collision estimation, and the disadvantage is that vehicles, pedestrians / non-motorized vehicles are moving The medium speed may change continuously, so that the possibility of collision is also constantly changing, resulting in inaccurate prediction results

[0004] (2). The vehicle, pedestrian / non-motor vehicle collision position is described as a collision at a fixed point. The disadvantage is that it cannot be represented solely by the collision point The change of the collision position due to the speed change of the vehicle, pedestrian / non-motor vehicle in motion

[0005] (3). Vehicles, pedestrians / non-motorized vehicles are abstracted into points for collision prediction. The disadvantage is that the length and width of the car will not be considered, and the result will also be affected. impact, predictions are not accurate enough

[0006]Comprehensive analysis: In vehicle-pedestrian collision prediction, there are always unstable factors such as speed change, speed magnitude, vehicle length and width, etc., and idealized prediction methods often There is a large error, and accident detection or early warning often involves the life safety of pedestrians, it is necessary to provide a highly fault-tolerant system

Images