Safety Prognostic Analysis in Traffic System
In this chapter, the safety region based theory is extended to the traffic system. The theory applicability will be testified and we will discuss the safety prognostic analysis in traffic system.
The first part of the work is to research the traffic operation risk analysis model based on safety region. Firstly, we propose the state feature extraction algorithm based on SFS-PCA. Secondly, based on the extracted state features, we use LSSVM to assess safety region margin of traffic operation and distinguish traffic operation state. Finally, we test proposed algorithm on case of the American Alameda I-880 highway.
The second part is to research on the traffic crash risk evaluation model based on reliability theory. Firstly, we select state variables applying CART. Secondly, we fit the distribution function of each variable based on reliability and calculate the joint distribution function of variables. Thirdly, we apply SVM approximately to estimate limit state function (LSF), and assess traffic operation risk using LSF. Finally, we also test proposed algorithm on case of the American Alameda I-880 highway.
- 7.A.M. Hasofer, N.C. Lind, Exact and invariant second-moment code format. J. Eng. Mech. 100(1), 111–121 (1974)Google Scholar
- 10.R. Yu, Q. Shi, M. Abdel-Aty. Feasibility of incorporating reliability analysis in traffic safety investigation, Transportation Research Record: Journal of Transportation Research Board, No. 2386, Transportation Research Board of the National Academies, Washington, DC., (2013), pp. 35–41Google Scholar