IRIDIA - Supplementary Information (ISSN: 2684-2041)

Supplementary material for the paper:

Show me what you want: Inverse reinforcement learning to automatically design robot swarms by demonstration

Ilyes Gharbi, Jonas Kuckling, David Garzón Ramos, and Mauro Birattari (September 2022)


Table of Contents
  1. Supplementary video
  2. Orchestra
  3. User demonstrations
  4. Convergence of behavior
  5. Generated behaviors
  6. Code
  7. Statistical data

Supplementary video

Orchestra

The software used for building the arena and provide demonstrations of the swarm robots final position in simulation is called Orchestra.

Opening menu of Orchestra

Opening menu of Orchestra

Arena building menu

Arena building menu

Demonstration menu

Demonstration menu

User demonstrations

Below are the five demonstrations per mission that were provided to Demo-Cho to design the control software.

Homing

Demonstration for Homing Demonstration for Homing Demonstration for Homing Demonstration for Homing Demonstration for Homing

AAC

Demonstration for AAC Demonstration for AAC Demonstration for AAC Demonstration for AAC Demonstration for AAC

SAC

Demonstration for SAC Demonstration for SAC Demonstration for SAC Demonstration for SAC Demonstration for SAC

CFA

Demonstration for CFA Demonstration for CFA Demonstration for CFA Demonstration for CFA Demonstration for CFA

Convergence of behavior

Below are the plots showing the convergence of the generated behavior towards the demonstrated one. The lower the value, the closer the two behaviors.

Homing

Convergence plot for Homing

AAC

Convergence plot for Homing

SAC

Convergence plot for Homing

CFA

Convergence plot for Homing

Generated behaviors

The videos of the generated behaviors are available for download.

Code

The code and installation instructions are available for download.

Statistical data

The results of all experimental runs are available for download.