An adaptive transport management approach using imitation learning

Abstract

The area of Intelligent Transport Systems has been critical in traffic management and intelligent systems for the past decades. In this paper, we introduce a novel approach to traffic management. We develop a process that that starts with the development of a "game" based upon different road networks that are used to gather data based upon user actions. The user’s decision directly affect how traffic light states change. This data is then passed to an Imitation Learning model that can observe actions and imitate the same decisions on a similar road network.