Emerging robot swarm traffic

We discuss traffic patterns generated by swarms of robots while commuting to and from a base station. The overall question is whether to explicitly organise the traffic or whether a certain regularity develops `naturally'. Human driven motorized traffic is rigidly structured in two lanes. However, army ants develop a three-lane pattern in their traffic, while human pedestrians generate a main trail and secondary trials in either direction. Our robot swarm approach is bottom-up: designing individual agents we first investigate the mathematics of cases occurring when applying the artificial potential field method to three 'perfect' robots. We show that traffic lane pattern will not be disturbed by the internal system of forces. Next, we define models of sensor designs to account for the practical fact that robots (and ants) have limited visibility and compare the sensor models in groups of three robots. In the final step we define layouts of a highway: an unbounded open space, a trail with surpassable edges and a hard defined (walled) highway. Having defined the preliminaries we run swarm simulations and look for emerging traffic patterns. Apparently, depending on the initial situation a variety of lane patterns occurs, however, high traffic densities do delay the emergence of traffic lanes considerably. Overall we conclude that regularities do emerge naturally and can be turned into an advantage to obtain efficient robot traffic.