Implementation of SLAM Gmapping and Extended Kalman Filter for Security Robot Navigation System

Abstract

Robot Security is a robot that is responsible for security as well as patrolling. When patrol automatically, the robot requires a navigation system. The robot also needs a mapping system that is used to make a map of the environment and as information on its location according to the map. The sensors used are wheel odometry and LiDAR. The wheel odometry system often slips which causes errors in reading the actual position of the robot. To fix this problem, a sensor fusion between the Inertial Measurement Unit (IMU) and wheel odometry is used. To combine these sensors, namely using the Extend Kalman Filter (EKF) which runs on the Robot Operating System (ROS) operating system. Mapping and navigation system testing, carried out using IMU sensors and without IMU, towards the 5 target points that have been made. In the test without IMU, the error of the robot reaching the target was (x = 45.86%, y = 54.595%, and = 56.63%). After adding the IMU sensor, the robot error has decreased to (x = 2.02%, y = 1.796%, and = 0.22%). In conclusion, the data combined from the IMU sensor and wheel odometry could minimize the existing slip errors.