Enhancing navigation accuracy of Turtlebot3 Burger mobile robot

This project aims to identify the initial covariance value of a ROS-based mobile robot, specifically the Turtlebot3 Burger. The basic navigation of the robot requires a significant amount of data and resources to process the output path. To address this challenge, the Kalman Filter algorithm is implemented in this robot, as it is widely used for mobile robot navigation and system integration. One crucial parameter for implementing the Kalman Filter is the covariance matrix, which needs to be determined. Understanding the specifications of the robot is essential for programming and operating it effectively. The system model of this robot is developed based on the kinematic model of a two-wheeled mobile robot. To execute this project, an experimental setup consisting of a laptop and a robot, serving as the ROS Master and ROS Slave respectively, is required. Furthermore, the project aims to comprehend the function and efficiency of the robot's performance, including the LiDAR sensor, Inertial Measurement Unit (IMU) sensor, and Odometry sensor. These sensors are mounted on the robot to achieve accurate localization. An indoor experiment was conducted to determine the covariance value. Different sources of sensor information are fused into a single representational format called sensor fusion. By using an extended Kalman filter (EKF), data from Odometry and IMU sensors were combined to estimate the position and orientation of the mobile robot. The value 0.014405 represent x, y and z for position covariance matrix and value 0.004974 represent roll, pitch and yaw for orientation covariance matrix. This identified covariance value will serve as the initial covariance matrix for the implementation of the Kalman Filter-based system using this robot. The experimental results indicate that the proposed method is suitable and practical for real-world applications.