Special Session 96: 

Road traffic simulation system using bird`s eye view images by image recognition.

mitsuhiro namekawa
Kaetsu University
Japan
Co-Author(s):    Taisuke Hori, Shuya Kanagawa
Abstract:
We predict traffic congestion to use the system simulation. To perform the system simulation, the movement of the car is described as a model. As a general method to making the model of automobiles, we use the characteristics of automobiles and our experience. However, we use this method, for example, it cannot cope with the difference in characteristics between narrow urban roads and wide suburban roads. In this research, we adopt a modeling method to take a bird`s-eye view of an intersection connecting roads and analyze it by machine learning. In this method, a model of a car is constructed statistically. So, we use this method, build a model close to reality. We will presentation about this method and system.