Multisensor on-the-fly localization:

Abstract

This paper presents an approach for localization using geometric features from a 360 • laser range finder and a monocular vision system. Its practicability under conditions of continuous localization during motion in real time (referred to as on-the-fly localization) is investigated in large-scale experiments. The features are infinite horizontal lines for the laser and vertical lines for the camera. They are extracted using physically well-grounded models for all sensors and passed to a Kalman filter for fusion and position estimation. Positioning accuracy close to subcentimeter has been achieved with an environment model requiring 30 bytes/m 2 . Already with a moderate number of matched features, the vision information was found to further increase this precision, particularly in the orientation. The results were obtained with a fully self-contained system where extensive tests with an overall length of more than 6.4 km and 150,000 localization cycles have been conducted. The final testbed for this localization system was the Computer 2000 event, an annual computer tradeshow in Lausanne, Switzerland, where during 4 days visitors could give high-level navigation commands to the robot via a web interface. This gave us the opportunity to obtain results on long-term reliability and verify the practicability of the approach under application-like conditions. Furthermore, general aspects and limitations of multisensor on-the-fly localization are discussed.

An Active Loop Closure Fast SLAM Method Base on Topology – Section Lines Cascading Map

“…(1) Preprocess laser range finder data. It is divided into different continuous points set [12] by the adaptive method. The formula is used to judge two adjacent dots continuity as follows:…”

Planning Under Sensing Uncertainty

“…The first one is solved by applying the action sequence (E, N). The symmetry problems vanish for Example 12.2, which can also be solved by the sequence (E, N) because (1,2,3) is the only sequence of positions that is consistent with the actions and compass readings.…”

“…Combinatorial localization with very little sensing is presented in [134]. For further reading on probabilistic localization, see [3,37,76,78,83,85,93,108,109,135,155,166,167,187]. In [179,180], localization uncertainty is expressed in terms of a sensor-uncertainty field, which is a derived I-space.…”

Técnicas de navegación para un robot móvil utilizando sistemas de razonamiento espacial

“…Como resultado, se han desarrollado diferentes métodos probabilísticos muy robustos, como por ejemplo técnicas de filtrado de Kalman [52,69] filtros de partículas [106], o algoritmos de esperanza-maximización [71], entre otros. A pesar de esta evolución tan considerable, los métodos que ofrecen una mayor fiabilidad están basados en datos sensoriales en bruto [20,105], o en características de bajo nivel, como líneas [7]. Esto es un claro indicador de que, a pesar de que existen técnicas muy avanzadas para realizar tareas de localización y construcción de mapas, los robots son todavía incapaces de extraer información simbólica del entorno con la misma facilidad y precisión que lo hacemos los humanos.…”

“…Un trabajo similar en el enfoque pero mucho más minucioso y detallado es el sistema Finale de Kosaka y Kak [58], que también confía en la transformada de Hough para la detección de líneas verticales, pero supone que existen modelos geométricos detallados en 3D del entorno. Arras y Tomatis [6,7] también utilizan un EKF para fusionar características verticales en imágenes monoculares con paredes detectadas con un escáner láser 2D.…”