Action Selection via Learning Behavior Patterns in Multi-Robot Systems

The RoboCup robot soccer Small Size League has been running since 1997 with many teams successfully competing and very effectively playing the games. Teams of five robots, with a combined autonomous centralized perception and control, and distributed actuation, move at high speeds in the field space, actuating a golf ball by passing and shooting it to aim at scoring goals. Most teams run their own pre-defined team strategies, unknown to the other teams, with flexible game-state dependent assignment of robot roles and positioning. However, in this fast-paced noisy real robot league, recognizing the opponent team strategies and accordingly adapting one's own play has proven to be a considerable challenge. In this work, we analyze logged data of real games gathered by the CMDragons team, and contribute several results in learning and responding to opponent strategies. We define episodes as segments of interest in the logged data, and introduce a representation that captures the spatial and temporal data of the multi-robot system as instances of geometrical trajectory curves. We then learn a model of the team strategies through a variant of agglomerative hierarchical clustering. Using the learned cluster model, we are able to classify a team behavior incrementally as it occurs. Finally, we define an algorithm that autonomously generates counter tactics, in a simulation based on the real logs, showing that it can recognize and respond to opponent strategies.

[1]  Ki-Joune Li,et al.  Spatio-temporal Similarity Analysis Between Trajectories on Road Networks , 2005, ER.

[2]  Manuela M. Veloso,et al.  Feature selection for activity recognition in multi-robot domains , 2008, AAAI 2008.

[3]  Junzhong Gu,et al.  A modified Hausdorff distance based algorithm for 2-dimensional spatial trajectory matching , 2010, 2010 5th International Conference on Computer Science & Education.

[4]  Peter Stone,et al.  Know Thine Enemy: A Champion RoboCup Coach Agent , 2006, AAAI.

[5]  Brett Browning,et al.  Plays as Effective Multiagent Plans Enabling Opponent-Adaptive Play Selection , 2004, ICAPS.

[6]  Manuela M. Veloso,et al.  Coach planning with opponent models for distributed execution , 2006, Autonomous Agents and Multi-Agent Systems.

[7]  Dimitrios Gunopulos,et al.  Discovering similar multidimensional trajectories , 2002, Proceedings 18th International Conference on Data Engineering.

[8]  Manuela M. Veloso,et al.  SSL-Vision: The Shared Vision System for the RoboCup Small Size League , 2009, RoboCup.

[9]  Mahesh Kumar,et al.  Clustering seasonality patterns in the presence of errors , 2002, KDD.

[10]  Günter Rote,et al.  Computing the Minimum Hausdorff Distance Between Two Point Sets on a Line Under Translation , 1991, Inf. Process. Lett..

[11]  Jae-Gil Lee,et al.  Incremental Clustering for Trajectories , 2010, DASFAA.

[12]  Brett Browning,et al.  STP: Skills, tactics, and plays for multi-robot control in adversarial environments , 2005 .

[13]  Ronald L. Rivest,et al.  Introduction to Algorithms , 1990 .

[14]  Manuela Veloso,et al.  CMDragons 2010 Team Description , 2010 .

[15]  David W. Aha,et al.  Improving Offensive Performance Through Opponent Modeling , 2009, AIIDE.