Traffic Warning and Vehicular Homogeneity Measurement Based on Multidimensional HMM Algorithm

Abstract

Traffic issue does result in serious social and economic problems in most of the countries. Therefore, developing an automatic analysis system to help improve the traffic management is a must. In this paper, a novel traffic analysis mechanism based on multidimensional hidden Markov model and the homogeneity measurement of the vehicular route and mutual activity of vehicles is developed. In this model, each vehicle is traced to measure its homogeneity. After then, the multidimensional hidden Markov model is applied for further analysis to determine whether the route and driving behavior of a vehicle is belonged to normal driving or abnormal/violation/aggressive driving. The analyzed information will then provide for the administrator of the traffic control center to take optimal corresponding action such as warning or enforcement. Experimental results demonstrate the feasibility and validity of our proposed mechanism. It can be applied as a traffic management system.