Research On Key Technology In RoboCup 3D Soccer Simulation

With the deeper development of Artificial Intelligence, Robotics and Multi-Agent-System, how to control and organize multiple robots to cooperate and accomplish complex tasks has became a primary topic in the field of robotics. Robot soccer simulation system is a typical Multi-Agent-System platform, it is involved in many research domains, such as artificial intelligence, MAS, network communication, robotic motion control and the path planning.This thesis has mainly face to the RoboCup 3D soccer simulation system, it is more complex and truly as to the 2D soccer simulation system. This thesis focuses on the following aspects:1. Get a deep understand of RoboCup 3D soccer simulation system, including its architecture, operation mechanism and some essential technologies about RoboCup 3D Server.2. Establish the soccer simulation system's world model, and complete the world model's renewal and forecast by using the kalman filter algorithm.3. Design five basic skills for the players, such as using the separate control method to improve the efficiency of the kicking rate; using the dichotomy to calculate the point of intercepting; Design the Dribble skill by using a combination of evolutionary Artificial Potential Field algorithm and A* algorithm .4. Research the high level strategies, and divide them into two conditions, one is the players with ball, another is the players without ball.5. Accomplish the development of the Client program in the platform of Linux by using C++ language; Demonstrate and optimize the Client program in RoboCup soccer simulation.