节点文献

类人足球机器人控制系统的底层设计及决策研究

Control System Bottom Design and Decision-making Research of Humanoid Soccer Robot

【作者】 陈永利

【导师】 刘国栋;

【作者基本信息】 江南大学 , 检测技术与自动化装置, 2008, 硕士

【摘要】 类人机器人是机器人研究领域的一个重要分支,代表了机器人的尖端技术,是机器人研究的最高境界。类人机器人足球比赛是伴随着类人机器人技术和分布式人工智能的发展而迅速兴起的一种高科技对抗活动,积极参与该项活动将对我国类人机器人事业的发展起重要的推动作用。控制系统是机器人的核心部分。类人足球机器人关节众多,控制电路复杂,而比赛对控制系统的稳定性、实时性、协调性、功耗及体积等方面都提出了很高的要求,因此,合理的控制结构和优化的控制电路将是机器人在赛场上表现良好的基础。硬件方面,本文首先从当前Robocup机器人足球比赛对类人机器人的功能需求和规则要求出发,为实验室研制的类人机器人机械本体量身定做了一套能够较好完成比赛任务的控制系统。该控制系统采用主从式结构,主板选用XScale PXA270处理器,主要完成实时视觉信息处理、无线网络通信、决策及任务规划等上层任务;从板选用AVR单片机ATmega16L作为从处理器,主要完成倾角传感器信息的采集及转换、对各个舵机的直接控制等下层任务。主从处理单元之间采用I2C总线进行通信。然后完成了控制系统具体电路的设计工作,主要按照性价比高、容易购买到、集成度高、能耗小的原则选择了器件,以满足机器人对控制系统体积和能耗的要求。软件方面,嵌入式操作系统是机器人控制系统重要的组成部分。嵌入式Linux是机器人操作系统较好的选择。本文采用2.6.9内核版本,其可抢占式内存可以使机器人对比赛过程中各种突发事件具有更快速的响应能力。完成了BootLoader的烧写、内核及根文件的移植工作后,成功启动并运行了嵌入式Linux。最后,对主板的主要功能模块进行了必要的测试。决策方面,机器人在赛场上的一切智能活动都基于它的决策系统。本文首先在控制系统底层硬件平台基础之上构建了机器人比赛的自主决策系统,然后对该决策系统进行分层设计,最后运用自由状态机理论设计并实现了机器人的自主进攻策略和防守策略。真实环境中的实验及比赛结果验证了该决策方法的有效性。

【Abstract】 Humanoid robot is an important branch of the robot research field, on behalf of the advanced technology of the robot, being the highest realm on the robot research.Humanoid robot soccer competitions are a kind of high-tech confrontational activities,its emergence goes along with the rapid development of humanoid robot technology and distributed artificial intelligence.Participating in the games Actively will play a important role in promoting the development of China’s humanoid robot.Control system is the core part of the robot.Humanoid soccer robot have many joints and complex control circuit.To control system,the competitions have higher requirements on its stability,real-time,coordination,power consumption,size and so on.Therefore,a reasonable control structure and optimized control circuit will be the guarantee for the robot’s good performance in the games.In terms of hardware, firstly,on the basis of the function needs and rule requirements of the current Robocup robot soccer competition to the humanoid robot,we designed a set of control system for our lab’s humanoid robot body.The control system used a master-slave architecture. XScale PXA270 was chosen as the motherboard processor,and tasks such as real-time visual information processing,wireless communications,decision-making and mission planning needed the motherboard to complete.The slave board selected ATmega16L as the processor, mainly completing angle sensor information collection and conversion,servo motor controlling tasks. Between motherboard and slave board,we used I2C bus communicating.Then ,we completed the specific circuit design work of control system.In order to meet the robot control system size and power consumption requirements,those devices and components that are cost-effective, easy to buy, high integration level,low energy consumption were selected.In terms of software, embedded operating system is an important part in robot control system.Embedded Linux is a better choice for the robot operating system.Using version 2.6.9 kernel, the preemptive memory can make the robot rapidly response to a variety of emergencies in the games.After completing BootLoader programming,Linux core and file document of root transplanting,we had started and run the embedded Linux successfully. Finally,all functional modules of motherboard were tested simply.In decision-making aspect, all intelligent activities of robot on the court are based on its decision-making system. First,the decision-making system of robot competition was constructed on the control system hardware platform.Then,the decision-making system was designed hierarchically.finally,using the FSM theory,we designed and realized robot’s offensive and defensive stategy. Experimental in the real environment and competition results had proved the effectiveness of the decision-making method.

  • 【网络出版投稿人】 江南大学
  • 【网络出版年期】2009年 03期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络