http://hdl.handle.net/10045/1730 2013-05-18T18:24:28Z http://hdl.handle.net/10045/14173 Título: Large scale environment partitioning in mobile robotics recognition tasks Autor/es: Bonev, Boyan; Cazorla Quevedo, Miguel Ángel Resumen: In this paper we present a scalable machine learning approach to mobile robots visual localization. The applicability of machine learning approaches is constrained by the complexity and size of the problem’s domain. Thus, dividing the problem becomes necessary and two essential questions arise: which partition set is optimal for the problem and how to integrate the separate results into a single solution. The novelty of this work is the use of Information Theory for partitioning high-dimensional data. In the presented experiments the domain of the problem is a large sequence of omnidirectional images, each one of them providing a high number of features. A robot which follows the same trajectory has to answer which is the most similar image from the sequence. The sequence is divided so that each partition is suitable for building a simple classifier. The partitions are established on the basis of the information divergence peaks among the images. Measuring the divergence has usually been considered unfeasible in high-dimensional data spaces. We overcome this problem by estimating the Jensen-Rényi divergence with an entropy approximation based on entropic spanning graphs. Finally, the responses of the different classifiers provide a multimodal hypothesis for each incoming image. As the robot is moving, a particle filter is used for attaining the convergence to a unimodal hypothesis. 2010-05-01T00:00:00Z http://hdl.handle.net/10045/13292 Título: Large scale egomotion and error analysis with visual features Autor/es: Cazorla Quevedo, Miguel Ángel; Viejo Hernando, Diego; Hernández Gutiérrez, Andrés; Nieto, Juan; Nebot, Eduardo Resumen: Several works deal with 3D data in SLAM problem but many of them are focused on short scale maps. In this paper, we propose a method that can be used for computing the 6DoF trajectory performed by a robot from the stereo images captured during a large scale trajectory. The method transforms robust 2D features extracted from the reference stereo images to the 3D space. These 3D features are then used for obtaining the correct robot movement. Both Sift and Surf methods for feature extraction have been used. Also, a comparison between our method and the results of the ICP algorithm have been performed. We have also made a study about errors in stereo cameras. 2010-01-01T00:00:00Z http://hdl.handle.net/10045/12599 Título: Robot navigation behaviors based on omnidirectional vision and information theory Autor/es: Bonev, Boyan; Cazorla Quevedo, Miguel Ángel; Escolano Ruiz, Francisco Resumen: In this work we present a reactive autonomous robot navigation system based only on omnidirectional vision. It does not rely on any prior knowledge about the environment apart from assuming a structured one, like indoor corridors or outdoor avenues. The direction of the corridor is estimated from the entropy analysis of a 1-D omnidirectional image. The 2-D omnidirectional image is analyzed for obstacle avoidance and for keeping a safety distance from the borders of the corridor. Both methods are non-metric and no 3-D information is needed. The system performs well with different resolutions and the catadioptric sensor needs no calibration. We present results from indoor and outdoor experiments. 2007-09-01T00:00:00Z http://hdl.handle.net/10045/2200 Título: Extraction and error modeling of 3D data: application to SLAM Autor/es: Viejo Hernando, Diego; Cazorla Quevedo, Miguel Ángel Resumen: We are interested in using artificial landmarks obtained by a stereo system not only in SLAM-like algorithms but also feature extraction, map building, and so on. Using a stereo camera we can extract planes and geometrical primitives like that. In order to use these primitives a perceptual model of landmarks is needed. In this paper we present our perceptual model and some points about its possible use are detailed. 2006-01-01T00:00:00Z