Supplementary material for the paper:

Behavior Trees as a Control Architecture in the Automatic Design of Robot Swarms

Jonas Kuckling1, Antoine Ligot1, Darko Bozhinoski1, Mauro Birattari1 (April 2018)

1IRIDIA, Université Libre de Bruxelles, Brussels, Belgium


Table of Contents
  1. Abstract
  2. Foraging
    1. Videos
    2. Controllers
    3. Results
  3. Aggregation
    1. Videos
    2. Controllers
    3. Results

Abstract

Previous researches have shown that automatically combining low-level behaviors into probabilistic finite state machines produces control software that crosses the reality gap satisfactorily. In this paper, we discuss the interest of adopting behavior trees as architecture for the control software of robot swarms. We introduce AutoMoDe-Maple: an automatic design method that combines preexisting modules into behavior trees. To highlight the potential of this control architecture, we present robot experiments in which we compare AutoMoDe-Maple with AutoMoDe-Chocolate---a modular approach that generates probabilistic finite state machines---and EvoStick---an evolutionary swarm robotics design method---on two missions.

Foraging

The arena contains two source areas (black circles) and a nest (white area). A light is placed behind the nest to help the robots to navigate. In this idealized version of foraging, a robot is deemed to retrieve an object when it enters a source and then the nest. The goal of the swarm is to retrieve as many objects as possible.

Videos

AutoMoDe-Maple

AutoMoDe-Chocolate

EvoStick

The videos of all experimental runs are available for download.

Controllers

Examples of controllers produced by AutoMoDe-Maple and AutoMoDe-Chocolate are displayed below. Images of all the controllers produced by these two methods are available for download.

AutoMoDe-Maple

AutoMoDe-Chocolate

Results

The performance of the three design methods are available for download.

Aggregation

The swarm must select one of the two black areas and aggregate there. The objective function is computed at the end of the experimental run, and is maximized when all robots are either on the left or the right area.

Videos

AutoMoDe-Maple

AutoMoDe-Chocolate

EvoStick

The videos of all experimental runs are available for download.

Controllers

Examples of controllers produced by AutoMoDe-Maple and AutoMoDe-Chocolate are displayed below. Images of all the controllers produced by these two methods are available for download.

AutoMoDe-Maple

AutoMoDe-Chocolate

Results

The performance of the three design methods are available for download.