There's a common approach to AI that pervades most of the methods developed in such settings. This is often called the classical approach. It's associated with the assumption that any intelligent agent (robot, etc) needs a memory in which it stores information about the current state of the world. As the world changes, the robot's perceptions will change with it. The robot must decode these perceptions and update the information in its memory. In effect, this information forms the robot's model of the world. The robot also has certain goals. By combining these with the information in memory, it can make plans. These consist of sequences of primitive movements, which are fed to a ``motor control'' unit and obeyed. This will become clearer when you see the PopBeast examples. Although the classical approach has come in for a lot of criticism during the last ten years, it was the dominant paradigm for the previous twenty-five, and it has left a large legacy of terminology. So it's definitely worth demonstrating.