W.M. Cai, X.Z. Zhao
Pages: 87-94
Abstract
In order to solve the problem of large prediction time and location errors in traditional traffic accident prediction methods, this paper proposes a traffic accident prediction method at intersections based on big data of the Internet of Things. Firstly, with the support of Internet of things big data, the potential function of intersection traffic accidents is constructed from the perspective of vehicle state and intersection parameter state. Secondly, the minimum value of vehicle state change before and after crossing the intersection is calculated, and the continuity of vehicle state change is analyzed by Markov chain. Finally, the continuity analysis results are substituted into the intersection traffic accident potential function to realize the prediction of traffic accidents. The test results show that compared with the traditional methods, the prediction error of the accident time and the accident location can be less than 0.45s and 0.05m respectively.
Keywords: internet of things big data; crossroads; traffic accident; vehicle status; weighted markov chain; potential function