In Chapter 16, we learned how to add a new mobile robot, the TortoiseBot, to ROS. We decided on topic APIs, built a complete Gazebo model, and used low-level velocity commands to drive it around in simulation. In this chapter, we’ll take the next big step by getting the TortoiseBot to navigate autonomously (in simulation). To get there, we’ll take a series of smaller steps:
Visualize and verify transform data.
Add a laser sensor.
Configure and incorporate the navigation stack.
rviz to localize the robot and send navigation goals.
Recalling the setup with which we ended the previous chapter, we can launch a simulation of the TortoiseBot like so:
user@hostname$ roslaunch tortoisebot tortoisebot.launch
That launch file will start a Gazebo simulation of the TortoiseBot in an empty
world. Now it’s time to use
rviz to visualize the state of the (simulated)
robot. Leaving Gazebo running, start
rviz in the usual way:
You might wonder why Gazebo and
rviz are separate programs. They look
pretty similar: both give you a 3D view of a robot and allow you to visualize
various aspects of the robot and its environment. They are separate programs
because they play very different roles: Gazebo simulates a robot, while
rviz visualizes a robot. Gazebo is a substitute for a real robot in a physical environment, computing the effects of forces and generating synthetic sensor data. While the 3D ...