Control >> Aggregation >> Cluster

Cluster Pattern Formation

An interesting example of self-organised clustering is observed among bark beetle larvae. These larvae lay pheromone while feeding and react to the presence of pheromone moving in the direction of the gradient of its concentration. The process is auto--catalytic, that is, an increasing density of larvae corresponds to an increasing concentration of pheromone and vice versa. The movement of larvae is given by two components: movement toward zones of higher concentration of pheromone and random walk. Another behavioural component is thigmotaxis, the tendency of maintaining a contact with neighbours. The cluster sizes and size distribution of them depends highly on the initial density and distribution of larvae.

The clustering behaviour of the s-bots is implemented in a way similar to that of the bark beetle larvae. The s-bots are attracted by sound and by other s-bots , and turn on their speaker with a certain probability. Sound and s-bot attraction work as the pheromone concentration and thigmotaxis, while the light emission corresponds to the pheromone laying.

In clustering behaviour grippers are never used. Speakers are turned on with a certain probability. Low sound emission frequency effects the time of cluster formation, but generally it takes long time. Using the probabilistic architecture, small clusters can disperse easily, especially if there is a bigger and more stable cluster nearby, which emits more sound and creates a stronger attraction to free s-bots.

initial positions	final positions

See the movie (MPEG, 11Mb)

Control >> Aggregation >> Cluster