Global Map Generation and SLAM using LiDAR and Stereo Camera for tracking motion of Mobile Robot

Abstract

One of the topics of greatest attention in mobile robotics is related to the location and mapping of a robot in a given environment and the other, associated with the selection of the devices or sensors necessary to acquire as much external information as possible for the generation of a global map. The purpose of this article is to propose the integration between a caterpillar-type land mobile robot, SLAM tasks with LiDAR devices and the use of stereo vision through the ZED camera for the generation of a 2D global map and the tracking of the movement of the mobile robot using the MATLAB® software. The experiment consists of performing different detection tests to determine distances and track the position of mobile robot in a structured environment indoors, to observe the behavior of the mobile platform and determine the error in the measurements. The results obtained show that the integrated devices satisfactorily fulfill the tasks established in controlled conditions and in indoor environments, obtaining error percentages lower than 1 and 4% for the case of the LiDAR and the ZED camera respectively. An alternative was developed that solves one of the most common problems of mobile robotics in recent years and, additionally, this solution allows the possibility of merging other types of sensors such as inertial systems, encoders, GPS, among others, in order to improve the applications in the area and the quality of the information acquired from abroad.

Proposal of Simultaneous Localization and Mapping for Mobile Robots Indoor Environments Using Petri Nets and Computer Vision

“…It is important to highlight that, instead of the aforementioned image capture method, other alternatives can be used such as a picamera connected to a raspberry [79,80], omnidirectional cameras [81] and mobile robots equipped with these or with other types of sensors [82][83][84] to detect the landmarks. If a more robust system is needed, it is even possible to use a pattern recognition algorithm to identify doors without the need for landmarks [85][86][87] and board this system on an AGV.…”

“…As for maps based on the geometry of the explored area, sensors such as sonars and LIDAR are used to collect data related to the structure and shape of the objects that make up the environment [82,90,91]. The use of LIDAR requires more time and processing power from robots when compared to the proposed method, in addition to higher financial costs.…”