Fernando J. Mendiburu, David Garzón Ramos, Marcos R. A. Morais, Antonio M. N. Lima, and Mauro Birattari (March 2022)
|
Table of Contents |
We present Mate, an automatic off-line design method specialized in the design of spatially-organizing behaviors for robot swarms. Mate belongs to the family of modular methods known as AutoMoDe. We introduce Mate to study the automatic design of collective behaviors for missions in which the swarm is subject to spatial distribution constrains. In this paper, we produce control software for three missions with specifications related to the distribution of the swarm in the environment. We conduct experiments in simulation and with a swarm of 20 e-puck robots. Alongside Mate, we also conduct experiments with two other automatic design methods: Chocolate--a state-of-the-art instance of AutoMoDe; and EvoSpace---a method based on neuro-evolution. Early studies conducted with existing modular design methods have shown their limitations in the design of spatially-organizing behaviors for robots that operate under spatial constrains. By introducing a specialized method like Mate, we expect to overcome these limitations. The aggregate results of our experiments show that Mate performs significantly better than Chocolate and EvoSpace in the missions we consider. We introduce Mate, an automatic off-line method for the design of control software for robot swarms. Mate is a method specialized in the design of spatially-organizing behaviors. It belongs to the family of automatic modular design methods known as AutoMoDe. We conceived Mate to study how relative positioning between robots can facilitate the realization of robot swarms that display different forms of spatial organization. Existing instances of AutoMoDe do not consider specific relative distancing between robots. With Mate, we enable the design of such collective behaviors by adding a behavioral module that allows robots to form hexagonal regular patterns. We compare Mate with two other automatic design methods: Chocolate and EvoSpace. Chocolate is a state-of-the-art instance of AutoMoDe—on which we base the development of Mate, and EvoSpace is an implementation of the neuro-evolutionary approach. We assess the control software produced by the three methods in three missions related to the spatial distribution of the swarm in the environment. We conduct experiments in simulation and with a swarm of 20 e-puck robots. Results show that Mate designs collective behaviors in which robot swarms perform the missions by using relative positioning. The aggregate results show that Mate performs significantly better than Chocolate and EvoSpace in the set of missions we consider.
Below we show demonstrative videos of collective behaviors designed by Mate, Chocolate and EvoSpace. We conceived three missions for our research: ANY-POINT CLOSENESS, NETWORKED COVERAGE, and CONDITIONAL COVERAGE.We conducted experiments in simulation and with a swarm of 20 e-puck robots.
The complete collection of videos of the experimental runs we performed with physical robots, is available to download here. Pictures of the final positioning of the robots for each experimental run is available to download here.
The control software produced by Mate, Chocolate, EvoSpace, and the experimental results are available for download here.
The source code is available as free and open-source. Otherwise indicated, the software is available under the MIT License in the following repositories:
(i) ARGoS3-AutoMoDe for the implementation of Mate: https://doi.org/10.5281/zenodo.5893277.
(ii)ARGoS3-NEAT for the implementation of EvoSpace: https://doi.org/10.5281/zenodo.4849517.
(iii) ARGoS3 for the ARGoS3 simulator: https://doi.org/10.5281/zenodo.4889111.
(iv) argos3-epuck for the ARGoS3 plugin to simulate and operate the e-puck: https://doi.org/10.5281/zenodo.4882714.
(v) demiurge-epuck-dao for the software interface that enables the operation of the e-puck with Mate, Chocolate, and EvoSpace: https://doi.org/10.5281/zenodo.5893406.
(vi) experiments-loop-functions for the implementations of ANY-POINT CLOSENESS, NETWORKED COVERAGE, and CONDITIONAL COVERAGE: https://doi.org/10.5281/zenodo.5893411.
(vii) irace for the implementation of Iterated F-race (GNU General Public License): https://doi.org/10.5281/zenodo.4888996.