Self-Organized Flocking with a Mobile Robot Swarm: a Novel Motion Control Method

by Eliseo Ferrante, Ali Emre Turgut, Cristián Huepe, Alessandro Stranieri, Carlo Pinciroli, and Marco Dorigo
May 2012



This page contains all supplementary information for the article Self-Organized Flocking with a Mobile Robot Swarm: a Novel Motion Control Method.

Table of Contents
  1. Abstract
  2. Videos
    1. Without informed robots
    2. With informed robots
    3. Simulation videos
  3. Complete simulation results
    1. Without informed robots
    2. With informed robots
  4. Complete real-robots results
    1. Without informed robots
    2. With informed robots

Abstract

In flocking, a swarm of robots moves cohesively in a common direction. Traditionally, flocking is realized using two main control rules: proximal control, which makes the robots stay together and uses only local range and bearing information of neighboring robots; and alignment control, which makes the robot align in a common direction and resorts to more elaborate sensing mechanisms to obtain the orientation of neighboring robots. So far, in robotics limited attention has been given to an important component of the flocking mechanism, that is, motion control, used to translate the output of these two control rules into robot motion.

In this paper, we propose a novel motion control method: MDMC. Through simulations and real robot experiments, we show that with MDMC flocking in a random direction is possible without the need of alignment control and of robots having a preferred direction of travel. MDMC has the advantage to be implementable on very simple robots that the capability to detect the orientation of their neighbors. Additionally, we compare MDMC with another motion control method used in robotics, showing that with MDMC swarms can travel further in a desired direction when a small proportion of robots is informed about this direction. Finally, we also systematically study flocking under various possible conditions: with or without alignment control, with or without informed robots, with MDMC or with the state of the art method.


Keywords:self-organized flocking, motion control, collective decision-making, self-organization, swarm intelligence, swarm robotics.



Videos

Real robot videos without informed robots

This video shows a group of eight robots performing flocking without informed robots and without alignment control. The robots just use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper. In the video, we see that the group is able to organize even in this simple setting. This experiment represent one out of ten runs used to gather quantitative data.

Download the video in AVI format.

This video shows a group of eight robots performing flocking without informed robots and using alignment control. The robots use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper. Additionally, they use the light sensor and the communication device of the range and bearing to perform alignment control, as explained in the paper. Our frame of reference is set in such a way that the direction of the x-axis is pointing on the left towards the light source, and we use the counter clockwise convention. In the video, we see that the group is able to organize faster using alignment control than without alignment control. This experiment represent one out of ten runs used to gather quantitative data.

Download the video in AVI format.

Real robot videos with informed robots

This video shows a group of eight robots performing flocking with 2 informed robots and without alignment control. The robots just use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper. Two robots are informed about the goal direction (-90.1 degres - up) and is leading the rest of the group that is non informed. Our frame of reference is set in such a way that the direction of the x-axis is pointing on the left towards the light source, and we use the counter clockwise convention. In the video, we see that the group is able to decently follow the goal direction. This experiment represent one out of ten runs used to gather quantitative data.

Download the video in AVI format.

This video shows a group of eight robots performing flocking with 2 informed robots and with alignment control. The robots use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper. Additionally, they use the light sensor and the communication device of the range and bearing to perform alignment control, as explained in the paper. Two robots are informed about the goal direction (-22.5 degres - up left) and is leading the rest of the group that is non informed. Our frame of reference is set in such a way that the direction of the x-axis is pointing on the left towards the light source, and we use the counter clockwise convention. In the video, we see that the group is not able to effectively follow the goal direction. This experiment represent one out of ten runs used to gather quantitative data.

Download the video in AVI format.

Simulation videos

This video shows an example simulation run obtained with the ARGoS simulator. The video shows a simulation of a swarm composed of N=50 robots performing parallel organized motion. No informed robots and no alignment control is present. The robots just use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper.

Download the video in AVI format.

This video shows an example simulation run obtained with the ARGoS simulator. The video shows a simulation of a swarm composed of N=20 robots performing a rotating motion. No informed robots and no alignment control is present. The robots just use the range and bearing of their neighbors to for a lattice and they move according to the MDMC method proposed in this paper. This type of dynamics can be observed only in few cases and with a specific combination of swarm size and noise value. A further investigation of this type of dynamics is ongoing in a follow-up study.

Download the video in AVI format.

Complete simulation results

The following figures show the complete set of results that were obtained in simulation.

Results with no informed robots

Transient behavior

Steady-state behavior

Results with no informed robots

Transient behavior - 1% informed robot

Transient behavior - 5% informed robot

Transient behavior - 10% informed robot

Transient behavior - 15% informed robot

Transient behavior - 20% informed robot

Steady-state behavior with respect to the swarm size - 1% informed robot

Steady-state behavior with respect to the swarm size - 5% informed robot

Steady-state behavior with respect to the swarm size - 10% informed robot

Steady-state behavior with respect to the swarm size - 15% informed robot

Steady-state behavior with respect to the swarm size - 20% informed robot

Steady-state behavior with respect to the proportion of informed robots - 10 robots

Steady-state behavior with respect to the proportion of informed robots - 50 robots

Steady-state behavior with respect to the proportion of informed robots - 100 robots

Steady-state behavior with respect to the proportion of informed robots - 500 robots

Steady-state behavior with respect to the proportion of informed robots - 1000 robots

Complete real robots results

The following figures show the complete set of results that were obtained in simulation.

Results without informed robots

Results without informed robots