Development of an EKF-based localization algorithm using compass sensor and LRF

TL;DR

This paper develops an EKF-based localization system for mobile robots that fuses odometry, compass, and laser range data to improve navigation accuracy and robustness.

Contribution

It introduces a sensor fusion model using an extended Kalman filter to combine heterogeneous sensor data for enhanced robot localization.

Findings

Localization accuracy is significantly improved.

Sensor fusion reduces uncertainty in position estimates.

System supports diverse navigation applications.

Abstract

This paper presents the implementation of a perceptual system for a mobile robot. The system is designed and installed with modern sensors and multi-point communication channels. The goal is to equip the robot with a high level of perception to support a wide range of navigating applications including Internet-based telecontrol, semi-autonomy, and autonomy. Due to uncertainties of acquiring data, a sensor fusion model is developed, in which heterogeneous measured data including odometry, compass heading and laser range is combined to get an optimal estimate in a statistical sense. The combination is carried out by an extended Kalman filter. Experimental results indicate that based on the system, the robot localization is enhanced and sufficient for navigation tasks.