Probabilistic Potential Field for Robot Formation

We have succeeded in developing a new method of potential field for robot formation, called Probabilistic Potential Field (PPF). The approach is targeted on distributed control of robot swarm because the potential field is individually established by each robot. There are 3 major novelties of this approach different from previously developed methods:
1) The attractive and repulsive forces have been developed based on communication only, not including perception.
2) The potential field is a probabilitic function, not pre-assumed based on diffirential equations .
3) The probabilistic function has been developed based on empirical measure of the robots' communication

We sucessfully demonstrated that, a swarm of mobile robots can form and maintain hexagonal and square lattice. Impressingly, we can achieve the formation transition from square to hexagonal lattices without predetermination. The alternative is only obtained if the diagonal robots are fixed with their position in the formation.

This result is documented in an scientific paper and we hope to publish it very soon!

A realtime performance of hexagonal formation:


Square formation maintainance:


and


The formation transition from square to hexagon:






Browser our youtube for more video clips
http://www.youtube.com/user/TheMoreThanOne