The ROS Framework

The Robot Operating System (ROS)^[1] is a set of software packages and tools that help out with the creation of robotic applications. ROS provides libraries that support a vast amount of functionality one could desire from a robot e.g. navigation and simultaneous localization and mapping (SLAM). These two features were used extensively within the project. ROS also provides tools for simulation and visualization. These tools were used in debugging the application and test if the desired behaviour had been reached before the application was deployed into a real-world scenario for further benchmarking. A ROS application is made up of several nodes in a network called a ROS graph, where each node is a process that performs computation. Nodes are meant to operate on a very granular scale and a ROS application will, therefore, comprise of many nodes. In the case with the TB3B, there is, for example, one node controlling the LiDAR, one controlling the motors to the wheels and one node performing localization. Nodes communicate with each other by publishing messages to topics which are then subscribed to by other nodes. A message can contain a range of different data types. Standard primitive types are supported as are arrays of them. Messages also come in the form of service calls where a node sends a request message to a node to receive a reply in the form of a response message. Topics are named buses which nodes send messages over. Topics generate messages even if there is no subscriber. Nodes do not know who they are communicating with, it works on the same principle as FM radio where the interested will tune themselves to the continuous broadcast to find the information they are interested in. This model is however not appropriate for request and reply interactions between nodes. A concept known as a service is used to solve this problem. These can often be treated as if they were remote procedure calls.

Links

[1] The ROS Framework