Razvoj algoritma za avtonomno vožnjo robota v ROS/GAZEBO okolju

diplomsko delo

Gregor Popič (Author), Aleš Belšak (Mentor), Suzana Uran (Mentor), Jurij Rakun (Co-mentor)

Abstract

Diplomska naloga povzema snovanje algoritmov za avtonomno vožnjo robota v različnih simulacijskih okoljih. Pri tem so uporabljeni različni senzorji, kot so LiDAR, IMU in inkrementalni ter absolutni dajalniki za odometrijo. Algoritem je zapisan v C++ programskem jeziku, ki je namenjen meta robotskemu operacijskemu sistemu ROS. Ob tem je uporabljeno simulacijsko okolje Gazebo, ki služi kot pripomoček, za naključno generiranje testnih okolij, ki so bila tudi del tekmovanja Field Robot Event 2021. To je sestavljeno iz več nalog, kjer prva predstavlja avtonomno vožnjo robotov po ukrivljenih vrstah, na koncu katerih je potrebno zaviti v naslednjo vrsto. Drugo nalogo pa sestavljajo ravne vrst, deloma nepopolne vrste, kjer mora robot prevoziti le določene vrste, po v naprej določenem zaporedju. Razvita algoritma sta bila testirana v simulacijskem okolju in ovrednotena v primerjavi z drugimi podobnimi rešitvami na tekmovanju poljskih robotov Field Robot Event 2021.

Keywords

avtonomna vožnja;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FS - Faculty of Mechanical Engineering
Publisher:	[G. Popič]
UDC:	681.5.017(043.2)
COBISS:	101188099
Views:	285
Downloads:	50
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	The development of an algorithm for autonomous robot driving in ros/gazebo environment
Secondary abstract:	The diploma thesis summarizes the development process of algorithms for autonomous driving in different simulated environments. Various sensors are used in the process, such as LiDAR, IMU or incremental and absolute encoders for odometry. The algorithm is written in C++ programming language for the meta-operating system ROS. In addition, the Gazebo simulation environment is used, which serves as a tool for randomly generating test environments, that were part of the Field Robot Event 2021. The event consists of several tasks, starting with basic navigation where the robot must autonomously drive between curved rows of maze plants and turn in to the next inrow space at the end. The second task is made up of straight rows, where the robot must travel only through certain inrow spaces, in a predetermined sequence. The developed algorithms were tested in a simulation environment and evaluated in comparison with other similar solutions at the Field Robot Event 2021.
Secondary keywords:	FarmBeast;Field Robot Event;Autonomous Driving ROS;Rviz;Gazebo;
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za strojništvo, Mehatronika
Pages:	VI, 38 str.
ID:	13228011