Towards Developing a Framework for Autonomous Electric Vehicles Using CARLA: A Validation Using the Deep Deterministic Policy Gradient Algorithm

As the world of transportation goes towards fundamental transformations, the urban mobility experiences the dawn of autonomous electric vehicles. Motivated by this upcoming future, we aim to develop a framework which will facilitate research around digital twins, autonomous agents and efficient electrified transportation. This effort is build around a two seat compact electric vehicle, which is currently getting modified to act as an autonomous platform. The current work has a twofold purpose. Initially to describe the simulated approach towards building a realistic model of the proposed vehicle, which will allow realistic experimentation. Subsequently, we validate our approach by implementing a two step procedure towards an autonomous agent, which will be able to follow a predefined path in an urban environment. Initially we use a standard path planner, namely the
$A$
* algorithm, and once we have the desired trajectory we train an autonomous agent to perform the navigation task. The whole process is integrated in CARLA simulator by using a tailored GYM environment.