| Summary: | This paper aims to investigate Robot Operating System (ROS) based environment modelling and mobile robot position estimation considering dead reckoning and uncertainties. A mobile robot movement is analyzed in a few environment conditions by using Extended Kalman Filter with ROS to identify and examined the mobile robot estimation performance on its surroundings. The heading angle and initial state covariance performance are assessed with different mobile robot movement. The paper is organized mainly to describe the results from both simulation and experiment using Extended Kalman Filter that consists of undetermined and unpredictable environment states. For experimental verification, a Turtlebot3 equipped with a 360-degree LiDAR and IMU is being applied to demonstrate the performance of estimation in a situation that has unknown uncertainties in several conditions. Both simulation and experimental results indicates that state covariance is converging lesser than the initial state covariance in any environmental cases which is in contrast with the literatures. Besides, it is also found that the mobile robot heading angle is important to be accurate at all times for better estimation results. © 2021 IEEE.
|
|---|