Laser Scan Matching in Robot Navigation

Abstract

In this study, Laser scan matching is implemented for simultaneous localization and mapping in robot navigation. Two algorithms are presented to accomplish the laser scan matching using normal distribution transform. Firstly, an algorithm is proposed to collect and store the laser scan data obtained from the robot scan. The second algorithm is to perform scan matching and map building. Using the data of good scan matches, a neural network is trained and tested. The experimental work is carried out on the Turtlebot robot model in Gazebo simulator using MATLAB software.