For this study, we selected a simple behavior of ``constant-distance following.'' The simulation was performed using the arena robot simulator, capable of simulating a number of Pioneer-like robots in a user-specified world. In the simulation, two robots repeatedly circle the central corridor shown in Figure 1.
One robot is marked as a leader and tries to circle the corridor while periodically emitting a radio beacon. The other robot (the follower) traverses the same known path through the environment while using the radio beacon to keep a constant distance from the leader.The design of the controllers was decomposed into two subtasks. First, each robot must be capable of autonomously and reliably navigating infinite circles around the corridor of the environment in Figure 1. Next, the follower must also attempt to infer its distance from the leader using the radio beacon in order to follow it at a constant distance. The design of each of these tasks is described below.