The first chapters of this book introduced many of the abstract ROS concepts used for communication between modules, such as topics, services, and actions. Then, the previous chapter introduced many of the sensing and actuation subsystems commonly found in modern robots. In this chapter, we will put these concepts together to create a robot that can wander around its environment. This might not sound terribly earth-shattering, but such a robot is actually capable of doing meaningful work: there is an entire class of tasks that are accomplished by driving across the environment. For example, many vacuuming or other floor-cleaning tasks can be accomplished by cleverly designed and carefully tuned algorithms where the robot, carrying its cleaning tool, traverses its environment somewhat randomly. The robot will eventually drive over all parts of the environment, completing its task.
In this chapter, we will go step by step through the process of writing minimalist ROS-based robot control software, including creating a ROS package and testing it in simulation.
First, let’s create the workspace directory tree, which we will place in ~/wanderbot_ws:
user@hostname$ mkdir -p ~/wanderbot_ws/src user@hostname$ cd ~wanderbot_ws/src user@hostname$ catkin_init_workspace
That’s it! Next, it’s just one more command to create a package in the new workspace. To create a package called
wanderbot that uses
rospy (the Python client for ROS) and a few standard ...