Nowadays the need for intelligent, self-driving autonomous cars are increasing. There are several problems which have to be solved in order to accomplish this, like localization, mapping, collision-free path planning, and path following.
This thesis presents a solution for the above-mentioned problems and shows a working system with a 4-wheeled RC car which is capable of autonomous navigation. The project is implemented in Robot Operating System. Some components of the system were adopted and integrated from previous research results at the Department of Automation and Applied Informatics.
In the thesis the hardware components selection, the low-level electronic circuit design, and the microcontroller firmware are shown. The theoretical background of several path planning algorithms for nonholonomic car-like vehicles, and a simple path following method are also presented. Technical details are given about working with the Robot Operating System and the Raspberry Pi. A high-level controller software is designed in ROS to integrate SLAM, path planning, and simple path following. A simulation environment is built to accomodate the developing process. The simulated robot and the real robot are fully interchangeable. At the end of the thesis some possibilities for future work are recommended.