A Vision-Based Odometer for Localization of Omnidirectional Indoor Robots

Abstract

In this paper we tackle the problem of indoor robot localization by using a vision-based approach. Specifically, we propose a visual odometer able to give back the relative pose of an omnidirectional automatic guided vehicle (AGV) that moves inside an indoor industrial environment. A monocular downward-looking camera having the optical axis nearly perpendicular to the ground floor, is used for collecting floor images. After a preliminary analysis of images aimed at detecting robust point features (keypoints) takes place, specific descriptors associated to the keypoints enable to match the detected points to their consecutive frames. A robust correspondence feature filter based on statistical and geometrical information is devised for rejecting those incorrect matchings, thus delivering better pose estimations. A camera pose compensation is further introduced for ensuring better positioning accuracy. The effectiveness of proposed methodology has been proven through several experiments, in laboratory as well as in an industrial setting. Both quantitative and qualitative evaluations have been made. Outcomes have shown that the method provides a final positioning percentage error of 0.21% on an average distance of 17.2 m. A longer run in an industrial context has provided comparable results (a percentage error of 0.94% after about 80 m). The average relative positioning error is about 3%, which is still in good agreement with current state of the art.

Pose Estimation of Omnidirectional Camera with Improved EPnP Algorithm

“…If N is determined, reference point coordinates can be calculated by Equation (11). We can obtain X vir in different virtual camera coordinates system by Equation (7). Then, we reproject X vir to the omnidirectional camera system to obtain the location X c i .…”

“…Recently, omnidirectional vision system has become a hot topic for researchers in some fields such as robot driving [6,7], augmented reality [8], and video surveillance [9][10][11]. An omnidirectional vision system provides a 360-degree panorama in the horizontal direction and is composed of a common CCD and mirror.…”

Suboptimal Omnidirectional Wheel Design and Implementation

“…Alternatively, omnidirectional mobile robots usually have four mecanum wheels [15] or three or four synchro-drive wheels [16] with individual control of the orientation and rotation speed. These omnidirectional configurations can Sensors 2021, 21, 865 2 of 17 perform transversal displacements but have limitations when performing combined or fast displacements [17]. These motion limitations do not exist when using three or four optimal omnidirectional wheels [18], but then the kinematic control becomes more complex, the wheel assembly requires the use of expensive machinery, and the motion system is prone to the generation of vibrations due to wheel discontinuities [19][20][21].…”

A Sensor Fusion Method for Pose Estimation of C-Legged Robots

“…This is a largely due to the difficulty of obtaining a robust method to analyze and compute the pose and orientation of the robot. Some developments were carried out to try to estimate the displacement of the robot with proprioceptive sensors [34][35][36] or with the use of external sensors [37][38][39][40][41][42]. However, none of the previous works have made a sensor fusion with an odometry algorithm to obtain a better and more robust pose estimation.…”