On Road Intelligent Vehicle Path Predication and Clustering using Machine Learning Approach

This paper presents a vehicle path predication and clustering using vehicle trajectory-based data. Vehicles provided by the GPS navigator, contributes in-vehicle driving path for drivers to move in the proper direction. In order to scale down the hazards on road, the vehicle GPS sensor gathers the information on the movement of vehicles. Machine learning has the great intelligence to identify the destined direction of vehicle at different intersection, therefore in our paper presents the path prediction method based on two ensemble learning algorithms, are being made use and they are random forests and AdaBoost for model training. This paper also presents the significance of vehicle trajectory in the task of collision avoidance, road hazard informing, scheduling and routing, and efficient method of movement in smart vehicle identification. The path prediction and clustering are completely based on road geometry and kinematic mechanism. The work also includes the future estimation of uncertainty of vehicles using a machine learning model and predicted trajectories are used to estimate the time to collision and collision warning along with the quality of service by routing and scheduling the vehicle routes.