An Architecture of Sensor Fusion for Spatial Location of Objects in Mobile Robotics

Abstract

Abstract. Each part of a mobile robot has particular aspects of its own, which must be integrated in order to successfully conclude a specific task. Among these parts, sensing enables to construct a representation of landmarks of the surroundings with the goal of supplying relevant information for the robot's navigation. The present work describes the architecture of a sensing system based on data fusion from a CMOS camera and distance sensors. The aim of the proposed architecture is the spatial location of objects on a soccer field. An SVM is used for both recognition and object location and the process of fusion is made by means of a fuzzy system, using a TSK model.

A Comprehensive Framework for Perception in Robotic Soccer

“…A version of the proposed framework using an SVM multi-classifier to recognize the objects in the soccer field (same team robot, the other team robot and ball) can be found in (Oliveira et al, 2005). To evaluate this system, Table 4 summarizes results over different illumination values, according to Robocup rules for F180 robot competition (Robocup, 2007).…”

“…To integrate these sensor data, a novel calibration method, based on a regression SVM, was developed and it has shown a robust mapping between the calibration map and the obtained values in camera space, with a low average error of 1.88 degrees. Also, the vision system was evaluated, and the new scheme, with an addition of a cascade of boost rejection and an SVM, has given better performance than in (Oliveira, 2005). The Adaboost classifier decreased the computation cost for the object recognition task and the SVM, used at the last stage of the cascade, reinforce the decisions taken by the Adaboost.…”

“…This new approach represents an advance with respect of our first implementation in (Oliveira et al, 2005).…”

Qualitative distances and qualitative description of images for indoor scene description and recognition in robotics

“…A more recent research trend in literature is integrating results from distance sensor fusion with images taken from a camera in order to extract knowledge from the environment by means of an XML dataset [Zivkovic et al, 2008], an ontology [Zender et al, 2008], symbolic/qualitative information [Oliveira et al, 2005], etc. This is the direction of our approach.…”

Fuzzy Distance Sensor Data Integration and Interpretation

“…Moreover, the semantic meaning of these qualitative names could be improved and related to others in the future by means of an ontology, as it has been previously done by the authors in their approach for generating ontology-based qualitative description of images 17 . A more recent research trend in literature is integrating results from distance sensor fusion with images taken from a camera in order to extract knowledge from the environment by means of an XML dataset 14 , an ontology 15 , symbolic/qualitative information 16 , etc. This is the direction of our approach.…”