Localization and Initialization Algorithms based on UWB, LiDARand Odometry for Robotic Applications with ROS Ecosystem

Abstract

This paper describes the initialization problem along with the localization problem over the Turtlebot3 and many more mobile robots.The least squares techniques and the squared range measurements obtained from ultra-wide band (UWB) sensors are used forcalculating the initial robot position. Then by exploiting the initial position, Light Detection and Ranging (LiDAR) scans and scanmatching technique have been proposed to find the initial heading. Thus, the autonomous pose initialization, which is an importantproblem in robotic applications, is solved. The Extended Kalman Filter, which fuses UWB range measurements, odometry andAdaptive Monte Carlo Localization (AMCL) pose information, is adopted to localize the robot during its trajectory. New moduleshave been implemented for Robot Operating Systems (ROS) for real and simulation environments and they are made to be opensource to enable wide-spread adoption. The simulation results have shown that the proposed method’s Root Mean Square Error(RMSE) is 3 cm and it’s almost twice better in accuracy than the benchmarked method.