Publication: Implementación de algoritmo de scan-matching basado en PSO
Loading...
Identifiers
Publication date
2015-07-07
Defense date
2015-07-07
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
Uno de los campos en los que se está centrando la robótica los últimos años es el desarrollo de nuevos algoritmos que permitan localizar el robot en un entorno con el objetivo de mejorar su autonomía.
Este problema ha llevado al desarrollo de algoritmos de SLAM (Simultaneous Localization and Mapping) que buscan localizar al robot mientras se construye un mapa de su entorno y a la vez se explora el mismo usando este mapa. Un aspecto relativo a esta tarea en particular es el problema de scan matching, que es el objeto principal de este trabajo. El problema de scan matching cosiste en estimar la relación métrica (en posición y orientación) entre pares de escaneos, utilizando la información proporcionada por estos. Este es un problema muy común en robots móviles debido a que a menudo la información nos es dada por sensores extereoceptivos con información de profundidad.
En este caso ha desarrollado una solución al Scan Matching o correlación de escaneos basada en algoritmo del PSO (Particle Swarm Optimization). Se utilizará además, las propiedades del color para optimizar el proceso. La utilización del color permitirá un ahorro en coste computacional.
Este nuevo algoritmo se ha probado en un entorno real tridimensional con escaneos proporcionados por el sensor Kinect y se ha conseguido resolver el problema de scan matching.
One of the aspects on which robotics has been focused in the last years is the development of new algorithms that allow the localization of a mobile robot in order to improve its autonomy. This problem has caused the development of new SLAM (Simultaneous Localization and Mapping) algorithms that try to track the robot’s pose while building a map of the environment and exploring it at the same time using this map. A particular aspect of this task is the scan matching problem, which is the main topic of this work. The scan matching problem consists of estimating the metric relation (position and orientation) between a pair of scans using the information provided by these scans. This is a very common problem in mobile robotics because the information about the environment that is going to be represented is often given by an extereoceptive sensor with depth information. A solution for the Scan Matching problem based on PSO (Particle Swarm Optimization) has been implemented. It also uses colour features in order to improve the algorithm performance. Using colour proprieties the computational cost will be reduced. This new algorithm has been tested in a 3D real environment with scans given by the Kinect sensor and it can solve the scan matching problem.
One of the aspects on which robotics has been focused in the last years is the development of new algorithms that allow the localization of a mobile robot in order to improve its autonomy. This problem has caused the development of new SLAM (Simultaneous Localization and Mapping) algorithms that try to track the robot’s pose while building a map of the environment and exploring it at the same time using this map. A particular aspect of this task is the scan matching problem, which is the main topic of this work. The scan matching problem consists of estimating the metric relation (position and orientation) between a pair of scans using the information provided by these scans. This is a very common problem in mobile robotics because the information about the environment that is going to be represented is often given by an extereoceptive sensor with depth information. A solution for the Scan Matching problem based on PSO (Particle Swarm Optimization) has been implemented. It also uses colour features in order to improve the algorithm performance. Using colour proprieties the computational cost will be reduced. This new algorithm has been tested in a 3D real environment with scans given by the Kinect sensor and it can solve the scan matching problem.
Description
Keywords
Inteligencia artificial, Proceso de imágenes, Scan matching, SLAM (Simultaneous Localization and Mapping, PSO (Particle Swarm Optimization), Algoritmos