Y. Liu, Z.A. Zhang, Z.L. Shang, Z. Wang

Pages: 113-124

Abstract
Accurate prediction of the geographical spatial distribution location of traffic accidents can provide drivers with richer traffic geographic information, thereby reducing the rate of traffic accidents. However, the existing methods for predicting the geographical spatial distribution location of traffic accidents have the problems of large error and long time. Therefore, this paper proposes a method for predicting the geographical spatial distribution location of traffic accidents based on the traffic flow gig data. First, collect traffic flow big data during the accident through highway detectors. Secondly, the collected data is processed for exception deletion, missing completion and geocode conversion. Finally, according to the geographical spatial distribution characteristics, the geographical spatial distribution of traffic accidents is predicted through Kalman filter. The experimental results show that this method predicts the geographical spatial distribution of traffic accidents based on actual test results, with good prediction effect and high prediction efficiency.
Keywords: big data of traffic flow; traffic accident; geographical spatial distribution; nearest neighbor; kalman filtering