L. Cao, S.L. Wei, A. Misao

Pages: 71-82

Abstract
In order to reduce the loss of traffic hazards, a traffic safety early warning method based on the information entropy model of accident data is proposed. Build the information entropy model of accident data, obtain the situation of traffic safety and judge the level of traffic safety. If it is a danger, release the traffic danger warning signal immediately and implement the warning. The dynamic traffic data fault diagnosis and repair based on wavelet analysis method is used to build traffic safety early warning index with the repaired traffic data, so as to realize traffic safety early warning. The results show that the average of rejection rate and false recognition rate is about 0.01, the average of data completeness is as high as 99.59%, the early warning accuracy is closer to 1, the average incidence of collision accidents is reduced by 43%, which can achieve high-precision traffic safety early warning.
Keywords: traffic safety; accident data; information entropy model; early warning; data restoration