Model-based and machine learning-based high-level controller for autonomous vehicle navigation: lane centering and obstacles avoidance

Abstract

Researchers have been attempting to make the car drive autonomously. Environment perception, together with safe guidance and control, is an important task and is one of the big challenges when developing this kind of system. Geometrical or physical-based models, machine learning-based models, and those based on a mixture of both models are the three types of navigation methods used to resolve this problem. The last method takes advantage of the learning capability of machine learning models and uses the safeness of geometric models in order to better perform the navigation task. This paper presents a hybrid autonomous navigation methodology, which takes advantage of the learning capability of machine learning and uses the safeness of the dynamic window approach geometric method. Using a single camera and a 2D lidar sensor, this method actuates as a high-level controller, where optimal vehicle velocities are found, then applied by a low-level controller. The final algorithm is validated in the CARLA Simulator environment, where the system proved to be capable to guide the vehicle in order to achieve the following tasks: lane keeping and obstacle avoidance.