| Abstract [eng] |
Towards the replacement of traditional vehicles by autonomous systems, the aim is to enable an autonomous robot or vehicle to move on the road in a simplified environment, without human assistance, by making the necessary decisions itself and without violating the rules. An autonomous vehicle should travel in accordance with the established rules for as long as possible without making a critical mistake. Before self-driving vehicles are yet widespread, the aim of this work is not to realize the movement of a real vehicle, but only to create a prototype of how this could happen if the environment were as simple and predictable as in simplified simulation. With such information, it can be adapted to real conditions. The main task of the work is for a robot controlled in a simplified environment to develop algorithms for position detection and robot control using artificial intelligence and mathematical models. It is planned to study the problems of the movement of autonomous vehicles, their environment and other factors related to controlling the vehicle. The developed system will be able to drive in the traffic lane without leaving the road by minimizing the driving error. While developing such a system, research is conducted that reveals the correlations between different control models and position detection algorithms. |
