Study on Simulation of Warning Method and Evaluation of Car-following for Intelligent Drive

Abstract

Car-following behavior is one of the main driving behaviors of drivers in road traffic environment, the simulation experiment for car-following early warning of intelligent driving can provide driver with safe and efficient verification before Real Vehicle Experiments. The focus of the research is how to detect car-following distance by LiDAR and evaluate the car-following warning method under the simulation environment. In this paper, the mathematical model of LiDAR was established in the simulation environment so that we can obtain the Time Headway changing to warn driver to accelerate or not in real time. Next in importance, in Unreal Engine 4 virtual scene, by using the special collision detection box to simultaneously detect the longitudinal and lateral relative positions of the ego vehicle and the front vehicle, we can complete the comprehensive evaluation of car-following behavior in both horizontal and vertical directions. Ultimately, to verify the validity and reliability of this method, hardware-in-the-loop simulation of car-following warning method was realized by connecting the driving simulator. As a consequence, by using the method of this paper, we provide a feasible technical scheme for intelligent driving vehicles in respect of multi-lane car-following warning.