Title: Ground landmark-based localisation for mobile robots using a calibrated camera

Authors: Lili Ma; Mohamed Alborati; Xiaohai Li

Addresses: Department of Computer Engineering Technology, CUNY – New York City College of Technology, Brooklyn, NY, USA ' Department of Computer Engineering Technology, CUNY – New York City College of Technology, Brooklyn, NY, USA ' Department of Computer Engineering Technology, CUNY – New York City College of Technology, Brooklyn, NY, USA

Abstract: This paper discusses a non-iterative vision-based localisation scheme for an autonomous mobile robot designed to work in parking lots. The robot is equipped with a single calibrated camera. Commonly-available features such as yellow-coloured road lines are used as visual landmarks. Given the orientation of a yellow line, the orientation of the robot is computed based on the detected information of the yellow line on the image plane. The robot's position is calculated by further assuming that the position of one point on the yellow line is also known. Besides theoretical analysis, we also experimented the proposed scheme on a low-cost physical robot in a parking lot. Two issues pertaining to real-time implementation of the scheme on the robot are addressed: determination of image size that balances processing speed and feature extraction accuracy; effect of feature extraction error on the localisation accuracy.

Keywords: vision-based localisation; visual servoing; ground landmark; calibrated camera.

DOI: 10.1504/IJCVR.2022.122592

International Journal of Computational Vision and Robotics, 2022 Vol.12 No.3, pp.299 - 323

Received: 30 Aug 2020
Accepted: 24 Apr 2021
Published online: 03 May 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article