Αυτόνομη ρομποτική πλοήγηση με χρήση τεχνητών δυναμικών πεδίων

The objective of this thesis is the development of an algorithm for autonomous robot navigation in a well-known environment. The method that was used was Artificial Potential Fields (APF), which is an online reactive navigation method that has been studied extensively over the years. After an overview of the basic theory of APF, its mathematical structure is presented and explained analytically. The development was done by using the Robot Operating System (ROS) and the testing occurred in a sophisticated simulation platform, the Gazebo. The robot that the algorithm was implemented to, was an Unmanned Aerial Vehicle (UAV), the Firefly, part of the Ascending Technologies research line. Moreover, a real-time 3D environment mapping application was constructed by using a Visual-Inertial (VI) sensor. Although the algorithm is limited due to the assumption of prior knowledge of the environment, it manages to safely reach the goal point by creating a collision-free path. It works fast and finds the optimal path in almost every configuration that it has been tested on. Additionally, in order to solve a pretty common problem of the APF method -the local minimum trap-, an indoors navigation algorithm was developed that manages to escape and plans an alternative path. Finally, the 3D mapping produces satisfying results, by mapping all the navigated area perfectly