A TIGHTLY COUPLED HEADING AND POSITIONING SYSTEM FOR INDOOR NAVIGATION OF MOBILE ROBOTS

Abstract

This study shows the effectiveness of data fusion techniques to achieve robust heading and position information in localization systems. The proposed system that has tightly coupled structure calculates heading and position of the mobile robot using the absolute and relative positioning subsystems. The relative positioning subsystem obtains heading and position information by using the odometry information and kinematic model of the robot. Absolute positioning subsystem calculates position and rough heading information using ultrasonic signals. In this study, firstly the robust heading information is calculated by combining absolute and relative heading with conventional Kalman Filter. The correction on the relative position measurement has been made by using this robust heading information. Finally, in order to obtain better positional information, an adaptive Kalman filter is applied for fusing the absolute position and the corrected relative position. In the experimental study, the positional accuracy and precision of the system is obtained as 63 mm and 86% (for positional error<100mm) respectively in the test environment. The proposed system gives more reliable, continuous, and less noisy heading and position information and is suitable for many tasks in indoor environments.

Keywords

• indoor positioning
• heading
• fusion
• adaptive kalman filter
• tightlycoupled system