In the current word, where efficiency, safety, and comfort is a priority, automatization, is becoming more and more widespread. The creation and design of self-driving cars, which can arrive to any destination without the help of the driver, is one of the most known topics in the profession. Although the researches are still in their infancy, there are already news of many successful experiments and vehicles, like Wayno, Google’s self-driving car. Perhaps the most basic task to achieve the self-driving car, is the necessity for the vehicle to be able to plan its own path, like getting out of a garage. After the planning process, the car must be able to traverse the path successfully despite the presence of environmental disturbances, like a slope, or a strong side-wind.
During my work, a path planning algorithm created for real city environment has been implemented, along with a controller, allowing a car to traverse it. As an introduction, this document presents the essence of common path planning used in robotics, and describe the external function library called Robot Operating System (ROS), which establishes the main structure of the implementation. After that the implemented path planning method is described, which is based on the work of the 2007 DARPA Urban Challenge’s victorious team. Later the method of establishment for a possible velocity profiler builder, which purses minimal time is shown. For the verification, a simple simulation method is introduced, and the robot car used for the testing is described, for the better understanding of the validation methodology. Finally, a Model Predictive Control (MPC) based control method is introduced, which enables the car to efficiently follow the path.