Analysis of strategies for autonomus robot navigation applied to radiation survey in tunnels
Year: 2023
Type: dissertation
Abstract: This Master Thesis is an analysis of different strategies for the autonomous navigation of wheeled robots in tunnels. The main purpose of the project is to prevent people from doing dangerous tasks in unknown environments, such as performing radiation surveys in tunnels. The thesis is divided into two big topics, which are the main branches needed for developing an autonomous navigation system. The first topic is related to localization of the robot inside the tunnel and mapping of it. By using ROS (Robot Operating System) as the basis of the developed software, SLAM (Simultaneous Localization and Mapping) technique is performed. When executing the radiation survey, it is very important to have precise information of the position of the robot, so it is possible to know the level of radiation in every point of the tunnel, and it is easier to localize areas where interventions need to be done. Additionally, through SLAM it is possible to obtain an UpToDate map of the accelerator every time a survey is performed. The second topic is related to different control techniques. In this thesis, different types of controllers are tested to determine which is the best one for a specific application. Those controllers differ in the way the robot moves and also the way the information received by sensors is processed. A second part is included in the control techniques, which is the development of a collision avoidance system. For ensuring a safe navigation along the tunnel, the robot must be able to avoid obstacles and planning the path towards the goal. This project has been tested in an accelerator located at CERN. The main goal for using autonomous navigation in their robots is to prevent people from radiation exposure when surveys are done at the beginning of a technical stops in the accelerators.
Author: Islán Morríñiga, María
Institution: Universidad Politécnica de Madrid
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: CERN
OpenAlex ID: W7063957776
