Design And Implementation Of Robocup Humanoid Robot Intelligent Control System

Humanoid robot is one of the most concerned types of robots.It possesses similar body structure and sensory system to human beings,has certain decision-making ability and can imitate some human behaviors.It represents the top level of robotics and has significant research value.RoboCup is an international organization that promotes the research of intelligent robots by proposing a series of standard questions.The humanoid robot football game is one of the subprojects that requires teams to design and produce fully autonomous soccer robots.Based on the competition rules of RoboCup's 2017 class KidSize group,this article designs and implements an intelligent control system based on the SEU-UniRobot robot,which enables the robot to have the ability to perceive,make decisions,and execute,and can automatically complete football matches.In this paper,the overall design of the control system is proposed,and each sub-module in the system is discussed in detail.Finally,a comprehensive test of the robot intelligent control system is designed.This paper proposes a hierarchical control system design scheme.The bottom layer is responsible for the hardware-driven and real-time control of robot joints,and the upper layer is responsible for sensing information,generating decisions,and generating actions.Among them,the design process of the upper-level control system is mainly introduced.It is divided into three parts: controller selection,development environment construction,and control software design.The upper control system software adopts a mixture of deliberation control and reflection control structure and is divided into three subsystems: perception,decision and execution.Perceptual system describes the external environment by constructing the world model.This paper proposes an abstract sensor frame based on the observer pattern,and establishes a monocular ranging system and a vision subsystem to process data.The construction of the world model includes the three steps of object model preprocessing,Monte Carlo Localization and world model updating.This paper designs the execution system as an execution framework with parallel capabilities and proposes static and dynamic action generation.The decision-making system is divided into two layers: an external decision-making layer and an internal decision-making layer.A multi-layered finite state machine is used to describe the decision-making process.This article also introduced a series of debugging tools.These debugging tools have functions such as remote control,online data tracking,and offline simulation,which can simplify the operation and significantly improve the development efficiency.Finally,this paper designs a number of experiments to test the function of the control system.The experimental results show that the robot has the ability to perceive,decide and execute,and can independently complete the football game.