摘要
为了实现机器人自主完成击打乒乓球的操作,构建了由双目视觉摄像头、嵌入式中控、六自由度机械臂、视觉处理计算机和监控计算机等组成的乒乓球机器人系统.灵巧臂采用关节模块化设计方法,共有6个自由度,集成有关节位置、电机位置、关节力矩、驱动电流及温度等多种传感器.采用模块化、多层次的手臂电气结构,基于FPGA(field program-mable gate arrays)实现关节的通讯、电机驱动、传感器信息采集与处理等功能.中控处理器实现击打策略、机械臂正逆运动学、轨迹规划等,采用Intel Core II双核嵌入式CPU作为控制核心,以FPGA作为通信逻辑控制芯片.采用系统辨识方法预测乒乓球的轨迹,通过采集多个点,有效地减少偶然误差,提高预测精度.以预测的轨迹方程、乒乓球机器人的工作空间为基础,确定击球点的位置、姿态及速度;在机械臂的任务规划中,采用多个点、多个运动过程约束手臂末端的运动路径,提高击球成功率.击球实验证明了机器人系统及控制策略的有效性.
To achieve autonomous table tennis playing,this paper presents a table tennis robot system which consists of binocular vision camera,embedded central control processor,6 DOF robot arm,visual processing computer and monitoring computer.Based on the modular design,the 6 DOF dexterous robot arm integrates joint torque sensor,joint position sensor,motor position sensor,drive current sensor,temperature sensor and so on.The multi-level electronic structure arm can realize the joint communication,BLDC motors control and data collection by a field programmable gate array（FPGA）.The hitting strategy,positive and inverse kinematics and trajectory planning are functioned in the central control processor which employs the Intel Core II dual-core embedded CPU as the control core and the FPGA as the communication logic control chip.The system identification method,which can effectively reduce the accidental error and improve the prediction accuracy by collecting multiple points during the ball moving,is applied to the table tennis trajectory prediction.Based on the predicted trajectory equation of ball and the working space of robot arm,the position, posture and velocity of the robot arm on hitting the ball can be determined.In the motion planning,the arm motion is constrained by multiple points and multiple segments to improve the successful rate of hitting table tennis.Finally,the effectiveness of the robot system and the hitting strategy are validated by experiments.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2013年第3期33-38,共6页
Journal of Harbin Institute of Technology
基金
国家高技术研究发展计划资助项目(2008AA04Z602)
哈工大优秀青年教师培养计划资助(HITQNJS.2008.010)
关键词
乒乓球机器人
灵巧臂
击球策略
运动路径
轨迹规划
robot for table tennis
dexterous robot arm
hitting strategy
motion trajectory
trajectory planning
