Research On Decision-making System For NAO In Standard Platform League Of RoboCup

Standard Platform League of RoboCup is a platform for distributed artificial intelligence research on multi-humanoid robots.It combines the theories and techniques of environmental perception,signal processing,wireless communication,decision-making of group awareness.With the development of artificial intelligence,the research on it has been widely concerned by many scholars.In this paper,designed the module-based hierarchical decision-making for the 5V5 SPL of RoboCup.At the same time,in order to adapt to the complex and changeable game environment,this paper proposes a control strategy which can take into account the game information for the statics selection and switching,in the framework of B_Human system.Combining with the characteristics of distributed control,make the robot to work independently.The key technologies are as follows:Firstly,system analysis and basic motion design.For the NAO frame and RoboCup SPL,the module-based hierarchical decision-making is proposed.In addition,to make the robot move,there is a method for special actions designing.Secondly,path planning.Using the bidirectional RRT algorithm,and modify thepath by the artificial potential field method.Not only to get the rapidly expend tree for global dynamic environment,but also ease the blindness and non-optimal problem of RRT path planner,in a certain extent.And finally achieve verification in the static and dynamic environment.Thirdly,formation selection.The selection of the formation is a complex qualitative evaluation problem.According to the human football coach's decision-making,we consider seven factors which related to the ball position,offensive and defensive ability of both teams,the ability of teams and the threat to the opponent goal.Use the fuzzy comprehensive evaluation method to analyze the formation.And simulated in Sim Robot and Matlab environment.Lastly,Role assignment and behavior control.With the dynamic formation selection results,the robot team gets the group formation and assign different subtasks for each robot according to the characteristics of each robot's ability.Firstly,this paper define the role and uses the bidding function to carry out the dynamic role assignment of the offensive and defensive roles in different Line-up.According to the finite state machine design behavior selection model,designed the tasks for the Striker shoot and the goalkeeper.Finally,Sim Robot is used to verify the completeness and Effectiveness of the state machines.