Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automatically designing robot swarms in environments populated by other robots: an experiment in robot shepherding

David Garzón Ramos, Mauro Birattari

TL;DR

This work addresses the challenge of designing robot swarms that operate in environments populated by other autonomous robots, using robot shepherding as a testbed where shepherds coordinate a larger set of pre programmed sheep. It applies two automatic design methods, Pis-ta-cchio (automatic modular design) and EvoCMY (neuroevolution), to generate shepherd control software within the RM 3.3 e-puck framework and evaluates them in ARGoS3 simulations across nine mission-scenario pairings. Results show that automatic design discovers mission-specific interaction strategies, notably color signaling combined with proximity cues, enabling effective coordination with heterogeneous sheep controllers; both methods outperform manual and random-walk baselines, with no consistent edge of one over the other. The findings demonstrate the generality of automatic design for heterogeneous swarms operating in populated environments and motivate future work on simultaneous design of multiple swarms and more dynamic or adversarial agents.

Abstract

Automatic design is a promising approach to realizing robot swarms. Given a mission to be performed by the swarm, an automatic method produces the required control software for the individual robots. Automatic design has concentrated on missions that a swarm can execute independently, interacting only with a static environment and without the involvement of other active entities. In this paper, we investigate the design of robot swarms that perform their mission by interacting with other robots that populate their environment. We frame our research within robot shepherding: the problem of using a small group of robots, the shepherds, to coordinate a relatively larger group, the sheep. In our study, the group of shepherds is the swarm that is automatically designed, and the sheep are pre-programmed robots that populate its environment. We use automatic modular design and neuroevolution to produce the control software for the swarm of shepherds to coordinate the sheep. We show that automatic design can leverage mission-specific interaction strategies to enable an effective coordination between the two groups.

Automatically designing robot swarms in environments populated by other robots: an experiment in robot shepherding

TL;DR

This work addresses the challenge of designing robot swarms that operate in environments populated by other autonomous robots, using robot shepherding as a testbed where shepherds coordinate a larger set of pre programmed sheep. It applies two automatic design methods, Pis-ta-cchio (automatic modular design) and EvoCMY (neuroevolution), to generate shepherd control software within the RM 3.3 e-puck framework and evaluates them in ARGoS3 simulations across nine mission-scenario pairings. Results show that automatic design discovers mission-specific interaction strategies, notably color signaling combined with proximity cues, enabling effective coordination with heterogeneous sheep controllers; both methods outperform manual and random-walk baselines, with no consistent edge of one over the other. The findings demonstrate the generality of automatic design for heterogeneous swarms operating in populated environments and motivate future work on simultaneous design of multiple swarms and more dynamic or adversarial agents.

Abstract

Automatic design is a promising approach to realizing robot swarms. Given a mission to be performed by the swarm, an automatic method produces the required control software for the individual robots. Automatic design has concentrated on missions that a swarm can execute independently, interacting only with a static environment and without the involvement of other active entities. In this paper, we investigate the design of robot swarms that perform their mission by interacting with other robots that populate their environment. We frame our research within robot shepherding: the problem of using a small group of robots, the shepherds, to coordinate a relatively larger group, the sheep. In our study, the group of shepherds is the swarm that is automatically designed, and the sheep are pre-programmed robots that populate its environment. We use automatic modular design and neuroevolution to produce the control software for the swarm of shepherds to coordinate the sheep. We show that automatic design can leverage mission-specific interaction strategies to enable an effective coordination between the two groups.
Paper Structure (23 sections, 3 figures, 3 tables)

This paper contains 23 sections, 3 figures, 3 tables.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. AUTOMATIC DESIGN METHODS
  4. Robot platform
  5. Pis-ta-cchio
  6. EvoCMY
  7. EXPERIMENTAL SETUP
  8. Sheep control software
  9. Missions
  10. Aggre-gation
  11. Dis-persion
  12. Herding
  13. Baseline
  14. C-Human
  15. R-Walk
  16. ...and 8 more sections

Figures (3)

  • Figure 1: Experimental arenas for Aggre-gation, Dis-persion, and Herding. The figure shows an example of the starting positions of five shepherds (cyan) and ten sheep (yellow) in each mission.
  • Figure 2: Results per mission and sheep control software. The plots show the score obtained in the nine experimental scenarios, ten observations per method and scenario. Results per mission are organized in rows. Results per sheep control software are organized in columns. Results per design method are presented with grayscale box-plots, R-Walk ( ), C-Human ( ), EvoCMY ( ), Pis-ta-cchio ( ).
  • Figure 3: Friedman test that aggregates the results obtained in the nine experimental scenarios. The plot shows the average rank of each method and its confidence interval.