warthogtdp2d

7/30/2019 WarthogTDP2D

1/4

Warthog Robotics Team Description Paper 2012

Felipe Taha SantAna, Nicholas Makita Fujimoto, Ana Paula Barbosa , Murilo Portela Ribeiro , Yago de Matos Dorea,Douglas Alencar

1 Av. Trabalhador Sancarlense, 400, 13566-590, So CarlosUniversity of So Paulo, School of Engineering of So Carlos, Department of Electrical Engineering

So Carlos, So Paulo, Brazil

{ [email protected],[email protected] , [email protected] , [email protected] ,[email protected] ,[email protected] }

Abstract This paper describes the Warthog Robotics 2D soccer simulation team. The team was runner-up at the Brazilian Robotics

Competition (CBR) 2011 and, previously under the name of GEARSIM, won the CBR 2009 and the Latin American RoboticsCompetition (LARC) 2010. In this paper is presented the current team research focus, which has been designed and implementedwithin this last year.

I. INTRODUCTION

The Warthog Robotics 2D soccer simulation team is abranch of the Warthog Robotics Team [1], which wascreated in 2011, after the merge of two previous roboticsgroups, GEAR and USPDroids, both competitors at theBrazilian Robotics Competition during several years.Besides the 2D simulation league, the team also participatesin two other categories of the Robocup, the 3D simulationleague and the SSL.

Our approach to the simulation 2D league is theresearch of decision making in dynamic multi-agentsystems. Based on the uncertainty and the subjectivecharacter of each simulation 2D game, and on [2], [3], [4],[6], [7], we have chosen a fuzzy system technique todetermine the team formation, merging both previous works[6], [7], so that both behavior and positioning of each agentswere affected. This method will be presented further on thispaper.

The Warthog 2D Team is based on Agent2D base sourcecode [5], due its nice and clear implementation of the low-level layer, and its easiness in developing intelligentapproaches in the high-level layer.

II. FUZZY LOGIC

The fuzzy systems, introduced by Zadeh in 1973, consistof approaching the human and machine decisions. Thismakes the machine not only takes exact decisions, like yes orno, but also take intermediate decision: perhaps.

The main characteristics of a fuzzy system are thefollowing:

They express inaccuracios and uncertainties;

They are systems based on linguistic rules;

The reasoning is performed by approximatemethods;

Conclusions are obtained in parallel;

Ability to approximate non-linear complex system.

III. FUZZY SYSTEMFORMATIONCHOOSER

In our previous works, fuzzy systems were implementedon the coach, and they could alter either the positioning ofthe agents [7], or its behavior [6]. In this new approach, thecoach uses both fuzzy systems, blends them and generate anew formation which affect both positioning and behavior,changing completely the attitude of the agents, creating awhole new dynamics to the game.

This new fuzzy system has four inputs and two outputs.These parameters are, respectively:

- Time - Number of game cycles (Figure 1);- Successful Attacks - Value in percentage (%) (Figure

2);- Successful Defenses - Value in percentage (%) (Figure

3);- Ball position Value of the X coordinate, where the

value 0 is the minimum value, and 105 the maximum value(Figure 4);

- Stress Level Represents the behavior to be adopted(Figure 5);

Tactical Formation Represents the position to beadopted (Figure 6).

Fig. 1 Total game time [6].
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]


2/4

Fig. 2 % Successful Attacks [6].

Fig. 3 % Successful Defenses [6].

Fig. 4 Ball Position [7]*.

Fig. 5 Stress Level [6].

Fig. 6 Tactical Formations [7].

*some parameters are not in English:Muito perto stands for Too close;

Perto stands for close;Longe stands for Far away.

Due to the mathematical complexity of calculating thisnew fuzzy system and the limitation of time to performcalculations between each game cycle, we have decided tocalculate each fuzzy system separately, and the merge thetwo results.

Comparing with the previous works, instead of havingonly three behaviors with one positioning, or three differentpositioning with only one behavior, there are nine possiblecombinations of formations, as represent on table 1.

Behavior\Formation 4-4-2 4-2-3-1 4-3-3

Stressed 442_S 4231_S 433_S

Normal 442_N 4231_N 433_N

Light 442_L 4231_L 433_L

Table 1 Behavior and formation

It is interesting to point out that both fuzzy systemsshares one input (time), and they are influenced by eachother. When the formation chosen is the offensive (433), theteam tends to attack more, make more shots at goal. Thisaffects the percentage of successful attacks, which affects thechoice of the behavior. With more attacks, the behaviorchosen tends to be light, which prioritizes a more prudentbehavior, with less wrong passes. This generates a moredefensive behavior, which causes the ball position (xcoordinate) to be smaller, influencing the choice of theformation.

IV. TEAM BEHAVIOR

A. STRESSED

Situation in which the player behaves more aggressivelyin relation to the opponent, committing more fouls. Thestressed behavior is seen in Figure 7, in which the numberthree player committed a foul, resulting in the punishmentwith a yellow card. The only goal of a stressed team is goalscoring, If you are in possession of the ball, passes are maderapids, and several long dribbling, ever advancing field. Ifthe opponent is in possession of the ball, marking presses

the opponents, too many fouls being committed. There is noconcern with any energy spent, because all actions areperformed with maximum intensity. With this behavior youget a method of prevention to avoid at all costs the advanceof the adversary.

B. NORMAL

Situation in which the player has a normal behaviorduring the match (Figure 8). The team performs thisbehavior with his actions balancing two conditions: shorterexecution action with the greatest energy saving. This formmakes the team faster than a defaulting team relaxed,

however, will not spend as much energy as a team stressed.His attitude already are bolder, with attempted passes indepth, and a marking tighter. This type of behavior raisessome individual plays.


3/4

C. LIGHT

Situation in which the player behaves more cautious,avoiding any kind of foul. A team relaxed, as shown inFigure 9, performing the movements with the minimumintensity necessary for successful action. The players movevery slowly, dont make faults and their attitudes are alwaysobjective and reliable. Passes are made to the teammate thatis no opponent around, not doing any solo run, and they runin the balls direction with minimum intensity necessary toarrive before the adversary. This behavior can be seen as a

bonus to the team.

Fig. 7: Stressed behavior

Fig. 8: Normal behavior

Fig. 9: Light behavior

V. RESULTS

The results were obtained playing ten matchesperformed as follows:

Ten games against the 2010 world championHELIOS Akiyama (2010b);

Ten games against the 2009 world championWrightEagle Chen (2010);

Ten games against the 2010 latin-americanchampion GEARSIM GEAR (2010).

WARTHOGSIM HELIOS

1 3 2

3 3 5

3 1 4

4 0 2

5 0 3

6 0 3

7 0 2

8 1 4

9 1 2

10 2 5

AVERAGE 0.93 2.80

WARTHOGSIM WrightEagle

1 0 1

2 4 3

3 1 1

4 0 0

5 2 1

6 3 2

7 0 2

8 3 1

9 3 1

10 2 1

AVERAGE 1.17 1.80

WARTHOGSIM WrightEagle

1 0 0

2 1 1

3 1 3

4 2 1

5 4 2

6 1 2

7 3 1

8 3 1

9 4 410 2 4

AVERAGE 2.10 1.83


4/4

VI. CONCLUSION

This paper briefly describes the Warthog Robotics 2DSimulation Team, its current efforts and research areas. Wehave merged two previous researches in a new fuzzy systemthat affects both behavior and positioning of the agents.Tests and improvements are still being performed, since theamount of coding needed to implement all the newbehaviors were considerably high, but the preliminaryresults are quite satisfactory.

Our plans for the future includes some reinforcementlearning techniques applied to the goalie, and a swarmapproach for the agents navigation.

ACKNOWLEDGMENT

We are grateful to all the project supporters, that includes

all students who worked with us, our sponsors, the

Engineering School of Sao Carlos, and the Institute of

Mathematics and Computer Science.

REFERENCES

[1] Ross, T.J.:Fuzzy Logic with Engineering Applications, John Wiley,(2004).

[2] Nakashima, T. and Takatani, M. and Udo, M. and Ishibuchi, H.:AnEvolutionary Apparoach for Strategy Learning in RoboCup Soccer,vol.2,pp.2023-2028, (2004).

[3] Nakashima, T. and Takatani, M. and Manabu, N. and Ishibuchi, H.:TheEffect of Using Match History on the Evolutionary of RoboCup SoccerTeam Strategies, Synmposium on Computational Intelligence andGames, (2006).

[4] Hidehisa Akiyama, Agent2D Simulation League Team,online, available at: http://rctools.sourceforge.jp/pukiwiki/,consulted on January 2010.

[5] Fraccaroli, ES and Carlson, PM, A Fuzzy Approach For Modeling TheTeam WarthogSim, (2011).

[6] Fraccaroli, ES and Carlson, PM, Time GEARSIM 2010 Da Categoria

Robocup Simulation 2D, (2010).

warthogtdp2d

Documents