Tutorial: SLAM in complex large-scale GNSS-denied environments
This tutorial will provide a hands-on approach to the basic concepts for implementing a perception system in ROS, focusing on the SLAM problem. The described methods will be programmed or analyzed in a remote virtual machine running Ubuntu 18 and ROS (attendants should have a computer with the Google Chrome browser installed and a stable internet connection). The tutorial will show how to setup different ROS packages designed for SLAM, and how to evaluate them using pre-recorded datasets. Additionally, the participants will be given a ROS-based SLAM package with several pending improvements that they will be asked to implement. The developed code will be written in C++, using mainly the ROS and Eigen libraries. We will use ROS Development Studio, an online ROS version that needs no installation. It works in a remote server and requires laptops with Chrome browser and internet connection.