Prediction of road users behaviour and dynamic motion control in ROS

An automated driving system is a combination of various components that are capable to perceive, make decisions and operate an automobile: everything is done by electronics and machinery, the human driver leaves all the responsibility to the system. This system introduces automation into road traffic and includes handling of the vehicle, destination and awareness of surroundings. The focus of this thesis is to implement a dynamic motion planning algorithm in order to control continuously a vehicle from a start to a goal position avoiding collisions with known obstacles, both static and dynamic (people on the streets, other cars, bikes and so on). As the title explains, the first goal is to predict the trajectory of other road users so that the second goal can be achieved: the vehicle is able to evaluate a plan to reach a goal. This plan takes into consideration the actual position of the obstacles and their future positions in order to understand if a certain path could lead to collision before even starting the movement, and so generating the best trajectory that ensures safety, comfort and efficiency. The software has been developed in ROS (Robot Operating System) which allows access to largest ecosystem and community currently existing in the field of robotics.