ROS Navigation Tuning Guide
Kaiyu Zheng · Studies in computational intelligence · 2021
The ROS navigation stack has become a crucial component for mobile robots using the ROS framework. It is powerful, yet requires careful fine tuning of parameters to optimize its performance on a given robot, a task that is not as simple as it looks and potentially time-consuming. This tutorial chapter presents a ROS navigation tuning guide for beginners that aims to serve as a reference for the “how” and “why” when setting the value of key parameters. The guide covers parameter settings for velocity and acceleration, global and local planners, costmaps, and the localization package amcl. It also discusses recovery behaviors as well as the dynamic reconfiguration feature of ROS. We assume that the reader has already set up the navigation stack and ready to optimize it. The material in this chapter originally appeared on ROS Tutorials since 2017 with experiments conducted on ROS indigo. Nevertheless, the ROS navigation stack has been mostly stable since indigo, while the ROS 2 navigation stack shares significant overlap in terms of packages and key parameters, with several innovations. Hence, we present this material in the context of the most recent release (ROS Noetic) at the time of writing. We discuss notable changes made in the ROS 2 navigation stack with respect to ROS Noetic. A video demonstration of the outcome following this guide is available at https://y2u.be/1-7GNtR6gVk .