Purpose
This paper aims to address a mobile robot localization system that avoids using a dedicated laser scanner, making it possible to reduce implementation costs and the robot’s size. The system has enough precision and robustness to meet the requirements of industrial environments.
Design/methodology/approach
Using an algorithm for artificial beacon detection combined with a Kalman Filter and an outlier rejection method, it was possible to enhance the precision and robustness of the overall localization system.
Findings
Usually, industrial automatic guide vehicles feature two kinds of lasers: one for navigation placed on top of the robot and another for obstacle detection (security lasers). Recently, security lasers extended their output data with obstacle distance (contours) and reflectivity. These new features made it possible to develop a novel localization system based on a security laser.
Research limitations/implications
Once the proposed methodology is completely validated, in the future, a scheme for global localization and failure detection should be addressed.
Practical implications
This paper presents a comparison between the presented approach and a commercial localization system for industry. The proposed algorithms were tested in an industrial application under realistic working conditions.
Social implications
The presented methodology represents a gain in the effective cost of the mobile robot platform, as it discards the need for a dedicated laser for localization purposes.
Originality/value
This paper presents a novel approach that benefits from the presence of a security laser on mobile robots (mandatory sensor when considering industrial applications), using it simultaneously with other sensors, not only to guarantee safety conditions during operation but also to locate the robot in the environment. This paper is also valuable because of the comparison made with a commercialized system, as well as the tests conducted in real industrial environments, which prove that the approach presented is suitable for working under these demanding conditions.
Map representation using hidden markov models for mobile robot localization