Maximum Likelihood Estimation Method for Speed Prediction of Vehicles in Mixed Traffic Condition

Overall objective of this study is to develop a traffic simulation model to realistically present the traffic flow in mixed traffic conditions. The simulation model can be used as applications in the fields of traffic operation and safety. This paper as a stage of the overall study presents a method to estimate speed of vehicles in mixed traffic condition in Ha Noi, Viet Nam. Several straight street segments in Ha Noi, Viet Nam were selected for observation. The traffic data includes vehicle types, speed that were extracted by using an image processing tool and other necessary data such as geometric conditions of streets. The result shows that there is a significant difference between free-flow speed of motorbikes and cars in different geometric conditions of the streets. The Maximum Likelihood Estimation method was used to develop a model for speed prediction of vehicles that takes into account types of vehicles and street geometric conditions. A Monte Carlo simulation method was also used to verify the developed speed model. It is concluded that the model can realistically represent the speed behavior of vehicles in the mixed traffic condition.