Aristotle University Of Thessaloniki (AUTH)
Ενοποιημένο σύστημα ενός ή πολλαπλών ρομπότ για την πλοήγηση και εξερεύνηση σε αστικό περιβάλλον με σκοπό τον εντοπισμό ανθρώπινης ζωής (USAR)
Abstract
dc:descriptionThere is a need of autonomy in many aspects of everyday life and especially in tasks that are hard to be performed by manpower or that are dangerous to the human life. One such task is Urban Search and Rescue (USAR) applications.The current Ph.D. focuses on the proposal of a unified solution for the autonomous exploration of ‚disaster environments‛ from one or multiple robotic vehicles. The term ‚disaster environments‛ is defined as civil constructions which have suffered structural deformations (for example due to an earthquake), such that traversing the space with a robotic vehicle is feasible. During the exploration, the robots check the surrounding environment for potential victims and pinpoint their position.A unified proposal for exploration and full coverage of an unknown a priori environment, from a single robot or a team of agents is addressed. The algorithms and the assumptions made, are focused on application in the international robotic competition RoboCup – RoboRescue, where some of them (Simultaneous Localization And Mapping and one-robot full exploration and coverage) have already been used successfully in 2009 (Austria) and 2011 (Turkey) from the robotic team P.A.N.D.O.R.A. of the Department of Electrical and Computer Engineering of Aristotle University of Thessaloniki.Innovations presented in the current Ph.D. are scattered in multiple fields of robotics. Initially, CRSM SLAM, a novel method for Simultaneous Localization And Mapping is proposed, using the concept of scan-matching, in order for the agent to perceive in detail its surrounding environment. Furthermore a survey of path planning methods is presented, which includes algorithms from different families, such as RRTs, Visibility Graphs, PRMs etc.. These methods are modified in order to to perform on a constant basis, regardless to the space morphology. Another contribution is in the way the robot chooses its next best target. The targets are exported as the set of the nodes of a topological graph, created by the space skeletonization using the Brushfire algorithm, in conjunction with morphological operations. After that, four different weights / costs are assigned to each possible target, they are combined efficiently and the next best goal is chosen. Finally the above methods are applied to a multi-agent system, where three different problems are addressed: a) exploration with known initial positions, b) with unknown initial positions and c) with unknown initial positions and problems in communication. The information merging is performed with spatial criteria, assuming that RFID tags are scattered in the environment, combining their topological information extracted by a localization module, RANSAC and ICP algorithms.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2013
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tsardoulias, Emmanouil
- Τσαρδούλιας, Εμμανουήλ
Subjects
dc:subject × 12- Πλοήγηση
- Εξερεύνηση και κάλυψη
- Πολλαπλά ρομπότ
- Εφαρμογή εντοπισμού και διάσωσης ανθρώπινης ζωής
- Navigation
- Exploration and coverage
- Multiple robots
- USAR (urban search and rescue)
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/39929