International Journal of Wisdom Based Computing, Vol. 1 (3), December 2011 123 A Dynamic Identification Method for Robotic Soccer Game Ayan Banerjee #1 , Ramanpreet Singh Arora *2 , Aruna Chakrabarty #3 #1 Electronics and Communication Engineering, #3 Computer Science Engineering, St. Thomas’ College of Engineering and Technology, 4, Diamond Harbour Road, Kidderpur,Kolkata-700034, Paschim Bengal, India, 1 ayanbanerjee.stcet@gmail.com, 3 arunachakraborty.stcet@gmail.com Applied Electronics and Instrumentation Engineering *2 Heritage Institute of Technology, Chowbaga Road, Anandpur, Kolkata-700107, Paschim Bengal, India. 2 ramanpreetsingh.hit@gmail.com Abstract—FIRA MIROSOT is a robo soccer game. It consists of small, middle and large league. Its large league consists of two teams of eleven players each. Each robot is classified in two ways – (i) its team tag, which indicates its team and (ii) a color tag which represents its identity in its team. Our paper proposes a new approach to identify a robot’s team identity and the identity of the robot itself in the team. Each robot wears a color jacket, the design of the color jacket is enables the visual system to effectively and efficiently identify the robot. The identification mechanism and the color jacket design is such that it identifies the robot with higher degree of accuracy, and reduced algorithm complexity and reduced execution time thereby increasing their reliability and effectiveness in a real time environment. It also allows using a large number of color tags. I. INTRODUCTION There has been a huge trend in research and development in the field of games in the last few decades. New technologies are invented for the purpose of intelligent and effective play of games. Hot technological topics like artificial intelligence and image processing are now embedded in the backbone of many games [2]. Robo Soccer is one such field where immense development is done by applying artificial intelligence and image processing [4],[5]. Robo Soccer is a game played by autonomous robots each reacting and responding to the play by applying their inbuilt artificial intelligence [4]. MIROSOT is such a game. It was opened in 1997.It is a model of the real world soccer game played by human players, where the robots are expected to show human like intelligence in the game play. Soccer is a dynamic game where things change fast and randomly so it can not be predicted ( i.e. the robots can not be pre-programmed ).Thus the robots using the concepts of Artificial Intelligence set their game plans and strategies during the play time according to the needs of the game[5],[6]. In the robosoccer architecture the automation part can be classified into three parts - 1. Vision System [8] 2. Decision Making Part 3. Motion Control However, the first and foremost concept that should be developed for a thorough model of the robosoccer is its Vision System. It is mandatory to identify all the robots of the home team and the individuality of each robot of the team. Thus for a good play the vision system should be accurate and reliable. It must be capable of giving the exact coordinates of the robots, identify the team robots and each member of the team and the ball position. It must be accurate in identifying the player and its algorithm should be easy so it can be modified in real time scenario if necessary [8]. In fast games, where every second counts, the visual data processing time must be small so that the robot can respond immediately. This paper deals with the vision system and the real world nature, hence we have put forth a novel approach to identify the distinct teams in the game and its individual players. II. COLOR TAG IDENTIFICATION METHODS For the purpose of detection of each robot (team and its identity in the team) various color jackets and designs were put forward in past [1][8]. The very basic design used in those days is displayed in figure 1. In this design, color tag 1 represents the team identity and region 2 represents the player identity. Since the region 2 represents is used to represent the players identity therefore all the eleven players must use eleven different color in region 2. The point O represents the centre of gravity (COG) of the robot and the arrow gives the forward direction of the robot. The disadvantages faced by the design were – to identify eleven different players there is a need for eleven different colors but in real time play due to inherent camera system noises and real-time illuminative variations many robots were identified wrongly. There was a problem in selecting the eleven different colors. In addition to this, calculating the COG of the robot was a lengthy and time consuming process.