Decision Making and Obstacle Avoidance for Soccer Robots

Robot-soccer game is characterized by a rapidly changing dynamic environment with moving obstacles. In this environment making a decision in a very short time, on-line adaptation and flexibility are very important parameters used in decision-making. Based on the decision made by a robot, the collision-free navigation of the mobile robot in this dynamic environment is also important. In the paper, the development of decision-making and navigation algorithms of a soccer robot are considered. Behavior-tree algorithm is presented and implemented for control of the mobile robots. At the same time, the type-2 fuzzy logic system is used for solving the obstacle avoidance problem. The use of the presented algorithm allows the decreasing of the path length of the robot. The efficiency of the presented algorithms has been tested with simulations and practical experiments. The obtained results demonstrate the suitability of using presented algorithms in the navigation of holonomic robots.