7.3 Conclusions
As experiments have shown the Max-sum algorithm applied over the proposed factor graph framework, where each robot has its own function and variable nodes, is very attractive for the coalition formation problem. Such procedure in a distributive manner but with a few messages exchange guarantees optimal solutions and by the periodic updating of the system neighborhood also the fault tolerance.
However this system, which relies on time-based utility functions, needed to be completed by a lower level of coordination, i.e. the robotics collision avoidance, which could make robots reach their chosen task without collisions and navigation faults. This is the reason we integrate our high level coordination with a kinodynamic but distributed collision avoidance system [3], we called cooperative collision avoidance system.
These systems are merged together thank to their common framework, the factor graph, and the distributed procedure, the Max-sum algorithm, however it was interesting to introduce and develop a theoretical architecture completely different from those studied in the state of the art, we called coordinated collision avoidance.
Such hybrid structure has permitted to trade off between greedy approaches and optimal solutions algorithms, making robots able to avoid collisions, rapidly choose tasks and optimizing those choices at the same time. However all carried out experiments are simulated within the Gazebo simulator over ROS middle-ware, hence future works could concern the tests and analysis of the proposed system on real Pioneer 3AT robots and real-life environments.
Acknowledgements
The author wishes to thank Michele Roncalli for giving some important features in order to develop the Max-sum algorithm and Nicolò Boscolo for helping the implementation of such distributing algorithm and giving a kinodynamic col-lision avoidance system. Moreover the author also wishes to thank Matteo Munaro and Stefano Micheletto for giving precious advise about ROS middle-ware.
Bibliography
[1] O. Shehory and S. Kraus. Methods for task allocation via agent coalition formation, Artificial Intelligence, Volume 101 (1–2), 1998, 165–200.
[2] T. Rahwan, S. D. Ramchurn, A. Giovannucci, V. D. Dang and N. R. Jennings. Anytime optimal coalition structure generation, in Proceeding of the 22nd conference on artificial intelligence, AAAI-07, Vancouver, Canada, 20AAAI-07, 1184–1190.
[3] N. Boscolo, R. De Battisti, M. Munaro, A. Farinelli and E. Pagello.
A distributed kinodynamic collision avoidance system under ROS, in Proceedings of 12th Int. Conference on Intelligent Autonomous Systems (IAS-12), JeJu Island, Korea, June 26-29, 2012.
[4] C. Candea, L. Iocchi, H. Hu, L. Iocchi, D. Nardi and M. Piag-gio. Coordination in Multi-Agent RoboCup Teams, Robotics and Autonomous Systems, 2001, Volume 36(2-3), 67-86.
[5] L. Iocchi, D. Nardi, M. Piaggio and A. Sgorbissa. Distributed Co-ordination in Heterogeneous Multi-Robot Systems, Autonomous Robots, 2003, Volume 15(2), 155-168.
[6] E. Pagello, A. D’Angelo, C. Ferrari, R. Polesel, R. Rosati and A. Speranzon. Emergent Behaviors of a Robot Team Performing Cooperative Tasks, Advanced Robotics, 2003, Volume 17(1), 3-19.
[7] E. Pagello, A. D’Angelo and E. Menegatti. Cooperation Issues and Distributed Sensing for Multi-Robot Systems, Proceedings of the IEEE, 2006, Volume 94(7), 1370-1383.
[8] L. Chaimowicz, R. Kumar and M. Campos. A Mechanism for Dynamic Coordination of Multiple Robots, Autonomous Robots, 2004, Volume 17(1), 7-21.
[9] A. Farinelli, G. Grisetti, L. Iocchi and D.Nardi. Coordination in dynamic environments with constraints on resources, IROSWK02, Dept. of Informatics and Systems, University "La Sapienza", 2002.
121
[10] O. Zweigle, R. Lafrenz, T. Buchheim, H. Rajaie, F. Schreiber and P. Levi. Cooperative Agent Behaviour Based on Special Interaction Nets, Intelligent Autonomous Systems 9, 2006, Volume 0, 651-659.
[11] A. Farinelli, H. Fujii, N. Tomoyasu, M. Takahashi, A. D’Angelo and E. Pagello. Cooperative control through objective achievement, Robotics and Autonomous Systems, 2010, Volume 58(7), 910-920.
[12] B. Gerkey and M. Matarić. On Role Allocation in RoboCup, Com-puter Science, 2004, Volume 3020/2004, 43-53.
[13] B. Gerkey and M. Matarić. Are (explicit) multi-robot coordination and multi-agent coordination really so different?, Proceedings of the AAAI Spring Symposium on Bridging the Multi-Agent and Multi-Robotic Research Gap, 2004, 1-3.
[14] B. Gerkey. On multi-robot task allocation, Technical Report CRES-03-012, University of Southern California, 2003.
[15] M. Matarić, G. Sukhatme, E. Ostergaard. Multi-Robot Task Al-location in Uncertain Environments Autonomous Robots, 2003, Volume 14(2-3), 255-263.
[13] M. Isik, F. Stulp, G. Mayer and H. Utz. Coordination without Negotiation in Teams of Heterogeneous Robots, Computer Science, 2007, Volume 4434/2007, 355-362.
[14] N. Lau, L. Lopes, G. Corrente, N. Filipe and R. Sequeira. Robot team coordination using dynamic role and positioning assignment and role based setplays, Mechatronics, 2010.
[15] F. R. Kschischang, B. J. Frey and H. Loeliger. Factor Graphs and the Sum-Product Algorithm, IEEE TRANSACTIONS ON IN-FORMATION THEORY, 1998, Volume 47, 498-519.
[16] S. M. Aji and R. J. McEliece. The generalized distributive law, IEEE Transactions on Information Theory, 2000, Volume 46(2), 325-343.
[17] S. D. Ramchurn, A. Farinelli, K. S. Macarthur and N. R. Jennings.
Decentralized Coordination in RoboCup Rescue, The Computer Journal, 2010, Volume 53(9), 1447-1461.
[18] B. J. Frey and D. Dueck. Clustering by passing messages between data points, Science 315(5814), 2007, 972–976.
[19] A. Farinelli, A. Rogers, A. Petcu and N.R. Jennings. Decentralized coordination of low-power embedded devices using the max-sum algorithm, Proceedings of the Seventh International Conference on Autonomous Agents and Multiagent Systems, 2008, 639–646.
BIBLIOGRAPHY 123 [20] A. Farinelli, A. Rogers and N. Jennings. Bounded Approximate De-centralized Coordination using the Max-Sum Algorithm, IJCAI-09 Workshop on Distributed Constraint Reasoning (DCR), Pasadena, California, USA, 46-59.
[21] O. Brock and O. Khatib. High-speed navigation using the global dynamic window approach, Robotics and Automation, 1999, Pro-ceedings 1999 IEEE International Conference, Volume 1, 1999, 341–346.
[22] J. Van Den Berg, S. Guy, M. Lin and D. Manocha. Reciprocal n-body collision avoidance, Robotics Research, 2011, 3–19.
[23] C. Clark, S. M. Rock and J. C. Latombe. Motion planning for multiple mobile robot systems using dynamic networks, IEEE Int.
Conference on Robotics and Automation, 2003, 4222–4227.
[24] T. Fraichard and H. Asama. Inevitable collision states. A step to-wards safer robots?, Intelligent Robots and Systems, 2003, (IROS 2003), Proceedings 2003 IEEE/RSJ International Conference on, Volume 1(1), 2003, 388–393.
[25] P. Hart, N. Nilsson and B. Raphael. A Formal Basis for the Heuris-tic Determination of Minimum Cost Paths, IEEE Transactions on Systems Science and Cybernetics, Volume 4(2), 1968, 100–107.
[26] J. Bruno, E. G. Coffman and R. Sethi. Scheduling independent tasks to reduce mean finishing time, Commun. ACM 17(7), 1974, 382-387.