Extended Kalman Filter Sensor Fusion in Practice for Mobile Robot Localization

Alaa Aldeen Housein,Yang Huang,Weiming Li,Gao Xingyu

doi:10.14569/ijacsa.2022.0130204

Alaa Aldeen Housein, Yang Huang + Show 2 more

Open Access

https://doi.org/10.14569/ijacsa.2022.0130204

Copy DOI

Abstract

Self-driving vehicles and autonomously guided robots could be very beneficial to today's civilization. However, the mobile robot's position must be accurately known, which referred as the localization with the task of tracking the dynamic position, in order for the robot to be active and useful. This paper presents a robot localization method with a known starting location by a real-time reconstructed environment model that represented as an occupancy grid map. The extended Kalman filter (EKF) is formulated as a nonlinear model-based estimator for fuse Odometry and a LIDAR range finder sensor. Because the occupancy grid map for the area is provided, just the inaccuracies of the LIDAR range finder will be considered. The experimental results on the “turtlebot” robot using robot operating system (ROS) show a significant improvement in the pose of the robot using the Kalman filter compared with sample Odometry. This paper also establishes the framework for using a Kalman filter for state estimation, providing all relevant mathematical equations for differential drive robot, this technique can be used to a variety of mobile robots.

Highlights

Mobile robots have been used to perform specialized tasks in a number of industries, including services, rescue, military, disaster relief, unmanned defense vehicles, and so on
The significance of this research rests in the framework it provides for fusing several sensors with a Kalman filter for robot localization, with the experimental results emphasizing notable reduction of errors in robot position
We focus on the problem of indoor robot localization, which involves determining the position of the robot x, y, and its orientation Ɵ

Summary

Introduction

Mobile robots have been used to perform specialized tasks in a number of industries, including services, rescue, military, disaster relief, unmanned defense vehicles, and so on. Researchers have added more sources in order to build a powerful localization approach [1].Odometry is one of the most important techniques to tackle the posture tracking problem, it uses encoder data to track the motion progress from a specified beginning position. This technique tracks motion from a known beginning position using encoder data, the encoded data is sent to the central processor, which uses a geometric equation to update the robot's position [2, 3]. The significance of this research rests in the framework it provides for fusing several sensors with a Kalman filter for robot localization, with the experimental results emphasizing notable reduction of errors in robot position.

Objectives

Discussion

Conclusion