Mobile Robot Localization Using Bar Codes as Artificial Landmarks

• Main Author: Ben-Hamid, Mahmud M. M.
• Format: Thesis
• Language: English; English
• Published: 2000
• Subjects: Mobile robots
• Online Access: http://psasir.upm.edu.my/id/eprint/10494/1/FK_2000_28.pdf; http://psasir.upm.edu.my/id/eprint/10494/
• Institution: Universiti Putra Malaysia

"Where am I' is the central question in mobile robot navigation. Robust and reliable localization are of vital importance for an autonomous mobile robot because the ability to constantly monitor its position in an unpredictable, unstructured, and dynamic environment is the essential prerequisite to build up and/or maintain environmental maps consistently and to perform path planning. Thus, selflocalization as precondition for goal-oriented behavior is a fundamental property an autonomous mobile robot needs to be equipped with. Accurate, flexible and low-cost localization are important issues for achieving autonomous and cooperative motions of mobile robots. Mobile robots usually perform self-localization by combining position estimates obtained from odometry or inertial navigation with external sensor data. The objective of the thesis is to present a pragmatic idea which utilizes a camera-based bar code recognition technique in order to support mobile robot localization In indoor environments. The idea is to further improve already existing localization capabilities, obtained from dead-reckoning, by furnishing relevant environmental spots such as doors, stairs, etc. with semantic information. In order to facilitate the detection of these landmarks the employment of bar codes is proposed. The important contribution of the thesis is the designing of two software programs. The first program is the bar code generation program which is able to generate five types of bar code labels that play a major role in the proposed localization method. The second program is the bar code recognition program that analyzes image files looking for a bar code label. If a label is found the program recognizes it and display both the information it contains and its coding type. Results concerning the generation of five types of bar code labels which are code 2 of 5, code 3 of9 , codabar code, code 128 and code 2 of 5 interleaved and the detection and identification of these labels from image files are obtained. In conclusion the thesis proposes a solution to mobile robot self-localization problem, which is the central significant for implementing an autonomous mobile robot, by utilizing a camera-based bar code recognition technique to support the basic localization capabilities obtained from a dead-reckoning method in an indoor environment.