Control >> Finding object/goal

Finding object/goal

In this section we describe the controller used for the s-bots to find an object/goal. As mentioned in the introduction, this is not a basic behaviour but one of the main behavioural capabilities that the s-bots should display to solve the scenario.

Unlike most of the other described controllers, we use a behaviour based architecture to control an s-bot. In general, we want to apply swarm intelligence methods, which take inspiration from social insect colonies, to allow a swarm of robots to solve complex tasks, while using simple control strategies for an individual robot.

In order to find an object/goal, the s-bots have to explore the environment. In ant colonies the problem of exploration and navigation is solved by establishing paths. This is done in a very simple and distributed manner. Ants lay trails of pheromone, a chemical substance that attracts other ants. Deneubourg et al showed that the process of laying a pheromone trail is a good strategy for finding the shortest path between a nest and a food source, thereby establishing a path that others can follow.

Inspired by this methodology of path establishment by pheromone laying, our approach to exploration is to use a chain of robots, a concept previously introduced by Goss et al, where the robots themselves act as trail markers, or beacons, in place of pheromone trails. We define a robotic chain to be a sequence of robots, where two neighbouring robots can sense each other and the distance between them never exceeds a certain maximum sensing range. In our case, the robots can visually sense each other by means of the omni-directional camera.

Chain formation

We try to reveal the general capabilities of a robot group performing chain formation by analysing the number of formed chains and their length when varying two probabilistic system parameters.

Goal search

Using the same controller that leads to chain formation, we discuss the results of a second set of experiments where a group of 10 robots has to find a prey object that is placed in the environment, and establish a connection between the prey object and a base station.

Control >> Finding object/goal