A Method of Driving Risk Prediction Based on Hidden Markov Model

Abstract

The present invention disclosed a method of driving risk prediction based on the Hidden Marcov model. First, based on the characteristics of vehicle operation, the risk of driving risks is divided through a clustering analysis method.The model estimates the impact of the driver's behavior and the surrounding traffic environmental characteristics on the probability of transferring the risk of driving;The initial value of the parameter of the state transfer probability matrix, establish a hidden Marcov chain model that reflects the evolution of driving status; 4. Get the characteristics of vehicle operation in real time, and based on the Hidden Marcov chain model, predict the future risk state in real time.The present invention establishes a hidden Marco model with a variable state transfer probability that can reflect the real -time change of the above characteristics, improves the accuracy and predictive accuracy of the driving risk model, and can meet the real -time requirements of anti -collision warning.

Description

Technical field [0001] The invention relates to the technical field of traffic safety evaluation and active safety of intelligent transportation systems, in particular to a method for predicting the degree of driving danger based on a hidden Markov model. Background technique [0002] The increase in the number of cars and the rapid development of the road transportation industry, while prospering the economy and facilitating people’s lives, have brought increasing pressure on the road traffic safety environment. The average annual number of deaths from traffic accidents in my country ranks first in the world, and road traffic accidents have become the field with the largest number of deaths in China's work safety. According to a number of research reports, if the driver can realize the danger of an accident 0.5s earlier and take corresponding corrective measures, 50% of accidents can be avoided; if it is 1s earlier, 90% of accidents can be avoided. Therefore, vehicle active saf...

AI Technical Summary

Problems solved by technology

In fact, the entire process from the formation of driving risks to the occurrence of dangerous conflicts is difficult to describe with a single early warning parameter, and more complex algorithms and models are needed for research

At the same time, the existing early warning model algorithms usually only consider the operating characteristics of the vehicle (such as speed characteristics and acceleration characteristics), while ignoring the impact of real-time driver behavior, road and environmental changes on the driving risk state, and cannot fully describe the driving state. The internal evolution law of time is not conducive to the accuracy and prediction accuracy of the driving risk model

Images