Abstract
Nowadays, mobile robots have extended to many different environments, where they have to move autonomously to fulfill an assigned task. With this aim, it is necessary that the robot builds a model of the environment and estimates its position using this model. These two problems are often faced simultaneously. This process is known as SLAM (Simultaneous Localization and Mapping) and is very common since when a robot begins moving in a previously unknown environment it must start generating a model from the scratch while it estimates its position simultaneously. This work is focused on the use of computer vision to solve this problem. The main objective is to develop and test an algorithm to solve the SLAM problem using two sources of information: (a) the global appearance of omnidirectional images captured by a camera mounted on the mobile robot and (b) the robot internal odometry. A hybrid metric-topological approach is proposed to solve the SLAM problem.
TopBackground
The SLAM problem has been extensively studied. Moravec and Elfes (1985) developed one of the first works in this area. They build a metric map by means of wide-angle sonar range measurements and a probabilistic approach. Subsequently laser sensors were introduced to improve the accuracy and computational efficiency of the algorithms. For example, Thrun (2001) develops a SLAM algorithm in which a team of robots builds a map online using laser sensors and a Monte Carlo approach. Lately, the use of cameras in the field of mobile robotics has become widespread due to the numerous advantages they offer (passive sensors, low cost, large amount of information, low power consumption, etc.). Many authors have studied the SLAM problem both using local features (Gil et al., 2010; Valiente et al., 2015) or global appearance (Paya et al., 2014; Berenguer et al. 2015).
Key Terms in this Chapter
Topological Map: It is a representation of the environment by means of a list of locations within a graph with connectivity relationships between them.
Metrical Map: It is a representation of the environment through geometrical information with certain accuracy.
Mobile Robot: It is an autonomous vehicle that is capable of movement in any given environment.
SLAM: It is the process of building a map of an environment while simultaneously the localization of the agent that compute the map is estimated.
Mapping: It is the creation of an internal representation of any given environment.
Omnidirectional Vision: It is a vision system that is capable of capturing all the information surrounding the system with a single image (360º).
Localization: It is the estimation of the position of an autonomous agent in a given map.
Gist: It is the meaning of a scene, or in other words, the spatial envelope of the scene.
Appearance Descriptor: It is a descriptor of an image that represents the global information of the same without extracting landmarks.
Probabilistic Localization: It is a localization task, where the information of all previous robot locations is used to estimate its current location.