L13: Mapping and Navigation with Nav2

Overview

This lecture covers how a mobile robot builds a model of its environment and uses that model to navigate autonomously. Maps provide the spatial representation a robot needs to plan paths, localize itself, and reason about the world. We focus on the occupancy grid, the metric representation used throughout Nav2, and on the standard map frame defined by REP 105. SLAM (Simultaneous Localization and Mapping) builds an occupancy grid from LiDAR data using slam_toolbox: scan matching estimates incremental motion, a pose graph stores the trajectory, and loop closure corrects accumulated drift. Once a map is saved, AMCL (Adaptive Monte Carlo Localization) localizes the robot against it using a particle filter. Finally, Nav2 orchestrates planning and control: global planners (NavFn, Smac) and local controllers (DWB, Regulated Pure Pursuit) operate on global and local costmaps to compute and execute collision-free paths. Goals are sent through the NavigateToPose action – either from RViz2 or programmatically through the nav2_simple_commander API. All hands-on examples use the rosbot_gazebo and nav_demo packages.

Learning Objectives

By the end of this lecture, you will be able to:

• Distinguish metric, topological, and semantic map representations.

• Explain the occupancy grid map representation and how LiDAR observations update it through Bayesian fusion.

• Build a map with slam_toolbox, save it with nav2_map_server, and reload it later for AMCL-based localization.

• Localize a robot against a known map using AMCL and a particle filter.

• Describe the role of global and local costmaps, the inflation layer, and the robot footprint.

• Distinguish global planners (NavFn, Smac Hybrid A*) from local controllers (DWB, Regulated Pure Pursuit).

• Explain how Nav2 uses a behavior tree to orchestrate planning, control, and recovery behaviors.

• Send navigation goals programmatically via the NavigateToPose action API using nav2_simple_commander.

Next Steps

• In the next lecture, we will cover Extra ROS Tools:

  • Foxglove Studio for live and recorded data inspection

  • ROS bags (ros2 bag record / play) for capture and replay

  • The RQT framework and its plugins

  • Final project status check

• Complete the exercises from this lecture before the next class.

• Read the Nav2 Configuration Guide and skim the Nav2 Behavior Trees documentation.