PROJECTPUBLICATIONSSWARM-BOTSprivate area
HARDWARESIMULATIONSCONTROL

Control >> Functional self-Assembly

Functional Self-Assembly

Consider a group of autonomous, mobile robots with the ability to physically connect to one another (self-assemble). The group is said to exhibit functional self-assembly if the robots can choose to self-assemble in response to the demands of their task and environment. We describe a robotic controller that enables s-bot robots to perform functional self-assembly.

Task

The task we consider requires a group of s-bots to navigate over an area of unknown terrain towards a target light source. If possible, the s-bots should navigate to the target independently. If, however, the terrain proves too difficult for a single s-bot, the s-bots should self-assemble into a swarm-bot group entity and collectively navigate to the target.

Figure 1: Scale diagram of the two experimental environments (view from above). Sbot starting positions are marked by crosses.

Controller

Figure 2: Behaviour transition model for the behaviour-based controller.

We use a distributed behaviour-based controller (see Figure 2). Each s-bot is fully autonomous. The same controller is executed on every s-bot. An s-bot starts by navigating independently towards the target light source. If the s-bot finds a hill too difficult for it to pass alone, or if it sees another s-bot that is either aggregating or assembled (sees blue or red), it illuminates its blue LEDs and starts aggregating. An aggregating s-bot can probabilistically trigger self-assembly by illuminating its red LEDs and becoming a static seed. Aggregating s-bots assemble to the seed s-bot or to already assembled s-bots (any red object). Assembled s-bots illuminate their red LEDs then perform group phototaxis once they can no longer detect any unassembled s-bots (can no longer see blue).

Results

We conducted a series of experiments in two different environments (see Figure 1) with groups of 1, 2 and 3 s-bots.

(a) (b) (c) (d) (e)

Figure 3: Example of a successful trial with 3 s-bots in Environment B. The s-bots start in a random configuration (a). One s-bot detects a slope it cannot overcome alone and activates blue LEDs (b). The other s-bots detect blue colour (local communication). The group aggregates and self-assembles (c,d). The s-bots collectively overcome the rough terrain and reach the target area (e,f)

Trials with 3 s-bot in Environment A

We conducted 20 trials. In every trial all 3 s-bot reached the target zone. In 19 out the 20 trials the s-bot correctly navigated independently to the target. In a single trial the s-bot self-assembled on the down slope of the hill and then performed collective phototaxis to the target. The incorrect decision to self-assemble was due to a colour misperception of a non-existent object by an s-bot.

Trials with a single s-bot in Environment B

We modified the controller to only execute Solo_Phototaxis behaviour. The s-bot was thus limited to navigating towards the target taking no account of the terrain encountered.

We conducted 20 trials. The s-bot failed to overcome the hill in 20 out of 20 trials. In each trial the s-bot reached the hill and then toppled backwards due to the steepness of the slope.

To confirm that the s-bot was failing due to the intrinsic properties of the slope, we repeated this experiment at a number of different constant speeds.

Trials with 2 s-bot in Environment B

We conducted 20 trials. The s-bot successfully detected the slope in every trial. Furthermore the s-bot always succeeded in assembling into a 2 s-bot swarm-bot. In 13 trials (65%) the swarm-bot succeeded in overcoming the hill. In the other 7 trials (35%) the assembled swarm-bot failed to overcome the hill. These failures happened when the assembled s-bot attempted to climb the hill in parallel.

Trials with 3 s-bot in Environment B.

We conducted 20 trials. The s-bot successfully detected the slope in every trial. In 16 trials (80%) all of the s-bot successfully self-assembled into a 3 s-bot swarm-bot. In each of these 16 trials the 3 s-bot swarm-bot went on to successfully reach the target area. Figure 3 shows a sequence of images from a typical trial.

In the remaining 4 trials (20%) the s-bot still managed in each case to self-assemble into a swarm-bot of 2 s-bots. In two of these 4 trials the swarm-bot went on to successfully reach the target area. In the two other trials the swarm-bot was obstructed by the third s-bot which failed to self-assemble.

Percentage of s-bots in Environment B trials succeeding for Self-assembly (A) and Completion of task (C).

Example movies:

References

  • O'Grady R., Groß R., Mondada F., Bonani M. and Dorigo M. Self-assembly on demand in a group of physical autonomous mobile robots navigating rough terrain, In Mathieu S. Capcarrere, Alex A. Freitas, Peter J. Bentley, Colin G. Johnson, Jon Timmis, editors, Advances in Artificial Life: 8th European Conference, ECAL 2005, volume 3630 of Lecture Notes in Computer Science, pages 272-281. Springer Verlag, Berlin, Germany, 2005


Control >> Functional self-Assembly

Swarm-bots project started
on October 1,2001
The project terminated
on March 31, 2005.
Last modified:
Fri, 27 Jun 2014 11:26:47 +0200
web administrator:
swarm-bots@iridia.ulb.ac.be