| Abstract [eng] |
This thesis presents the Development and Investigation of Navigation System for Mobile Robotic Platform. Recently, the navigation problems of mobile robots have received a lot of attention due to their increasing popularity in automation applications. Often in confined environments, traditional means such as GPS-based robot positioning are not sufficient. SLAM techniques are increasingly being used to solve such problems. The aim of this work is to find the most suitable algorithm from the analyzed localization and mapping methods by conducting experiments with different localization and mapping methods and parameters, so that the robot is able to carry out error-free navigation in dynamically changing environments and to transport loads. The analytical part provides an overview of the state-of-the-art localization and mapping algorithms in use. Applications of Hector SLAM, Google Cartographer, GMapping and Karto SLAM are analyzed, as well as the accuracy of the results obtained in the literature. In the methodological part, the design of the mobile platform is carried out using the YDLIDAR X4 laser range sensor and odometry data obtained from the mobile platform encoders. The Hector SLAM algorithm is tuned to obtain the best parameter values for robot localization and mapping. Navigation experiments are also performed using the move_base module. In the experimental part, the Google Cartographer, Hector SLAM, GMapping and Karto SLAM methods are applied and investigated. The accuracy of the localization and the map produced by each algorithm is considered, the performance is evaluated, and the most accurate algorithm for navigation is selected based on the experimental results. |
