Evolving a Repertoire of Controllers for a Multi-function Swarm

Automated design of swarm behaviors with a top-down approach is a challenging research question that has not yet been fully addressed in the robotic swarm literature. This paper seeks to explore the possibility of using an evolutionary algorithm to evolve, rather than hand code, a wide repertoire of behavior primitives enabling more effective control of a large group or swarm of unmanned systems. We use the MAP-elites algorithm to generate a repertoire of controllers with varying abilities and behaviors allowing the swarm to adapt to user-defined preferences by selection of a new appropriate controller. To test the proposed method we examine two example applications: perimeter surveillance and network creation. Perimeter surveillance require agents to explore, while network creation requires them to disperse without losing connectivity. These are distinct application that have drastically different requirements on agent behavior, and are a good benchmark for our swarm controller optimization framework. We show a performance comparison between a simple weighted controller and a parametric controller. Evolving controllers allows for specifying desired behaviors top-down, in terms of objectives to solve, rather than bottom-up.

[1]  Tucker R. Balch,et al.  Behavior-based formation control for multirobot teams , 1998, IEEE Trans. Robotics Autom..

[2]  Marco Dorigo,et al.  Towards group transport by swarms of robots , 2009, Int. J. Bio Inspired Comput..

[3]  Stéphane Doncieux,et al.  Incremental Evolution of Animats' Behaviors as a Multi-objective Optimization , 2008, SAB.

[4]  Anders Lyhne Christensen,et al.  Evolution of Hybrid Robotic Controllers for Complex Tasks , 2015, J. Intell. Robotic Syst..

[5]  Minoru Asada,et al.  Incremental Coevolution With Competitive and Cooperative Tasks in a Multirobot Environment , 2006, Proceedings of the IEEE.

[6]  Min Cheol Lee,et al.  Obstacle avoidance for mobile robots using artificial potential field approach with simulated annealing , 2001, ISIE 2001. 2001 IEEE International Symposium on Industrial Electronics Proceedings (Cat. No.01TH8570).

[7]  Anders Lyhne Christensen,et al.  Evolution of Collective Behaviors for a Real Swarm of Aquatic Surface Robots , 2015, PloS one.

[8]  Min Cheol Lee,et al.  Artificial potential field based path planning for mobile robots using a virtual obstacle concept , 2003, Proceedings 2003 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM 2003).

[9]  Anders Lyhne Christensen,et al.  Hybrid Control for Large Swarms of Aquatic Drones , 2014, ALIFE.

[10]  Stefano Carpin,et al.  Deploying teams of heterogeneous UAVs in cooperative two-level surveillance missions , 2015, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[11]  Luca Maria Gambardella,et al.  Cooperative self-organization in a heterogeneous swarm robotic system , 2010, GECCO '10.

[12]  Risto Miikkulainen,et al.  Evolving multimodal behavior with modular neural networks in Ms. Pac-Man , 2014, GECCO.

[13]  Kay Chen Tan,et al.  Evolutionary artificial potential fields and their application in real time robot path planning , 2000, Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512).

[14]  Jean-Baptiste Mouret,et al.  Neural Modularity Helps Organisms Evolve to Learn New Skills without Forgetting Old Skills , 2015, PLoS Comput. Biol..

[15]  Spring Berman,et al.  Optimized Stochastic Policies for Task Allocation in Swarms of Robots , 2009, IEEE Transactions on Robotics.

[16]  Sándor P. Fekete,et al.  Distributed cohesive control for robot swarms: Maintaining good connectivity in the presence of exterior forces , 2015, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[17]  D. Floreano,et al.  The evolution of information suppression in communicating robots with conflicting interests , 2009, Proceedings of the National Academy of Sciences.

[18]  Antoine Cully,et al.  Robots that can adapt like animals , 2014, Nature.

[19]  Jean-Baptiste Mouret,et al.  Illuminating search spaces by mapping elites , 2015, ArXiv.

[20]  Suranga Hettiarachchi,et al.  An Overview of Physicomimetics , 2004, Swarm Robotics.

[21]  Anders Lyhne Christensen,et al.  Hierarchical evolution of robotic controllers for complex tasks , 2012, 2012 IEEE International Conference on Development and Learning and Epigenetic Robotics (ICDL).

[22]  Dario Floreano,et al.  Evolved swarming without positioning information: an application in aerial communication relay , 2009, Auton. Robots.

[23]  Antoine Cully,et al.  How do Different Encodings Influence the Performance of the MAP-Elites Algorithm? , 2016, GECCO.

[24]  Risto Miikkulainen,et al.  Evolving Multimodal Networks for Multitask Games , 2012, IEEE Transactions on Computational Intelligence and AI in Games.