CASIA OpenIR  > 毕业生  > 博士学位论文
直角坐标机器人的运动规划及控制策略研究
其他题名Research on Motion Planning and Control Strategy for Cartesian Robots
杨平
学位类型工学博士
导师徐德
2011-05-24
学位授予单位中国科学院研究生院
学位授予地点中国科学院自动化研究所
学位专业控制理论与控制工程
关键词直角坐标机器人 乒乓球机器人 飞行轨迹预测 运动规划 运动控制 宏微协调运动 坐标对准 坐标纠正 Cartesian Robot Table Tennis Robot Ball Flying Trajectory Prediction Motion Planning Motion Control Macro-micro Coordinated Movement Coordinate Alignment Coordinate Correction
摘要直角坐标机器人具有本体结构简单,控制系统容易实现,可扩展性强等特点,广泛应用于工业现场。本文针对直角坐标机器人在视觉引导下的快速运动、大范围高精度跟踪以及高精度快速运动等问题展开相关研究工作。以五自由度乒乓球机器人为背景,研究在高速视觉引导下机器人与人对打乒乓球的运动规划和运动控制方法;以宏微机器人为背景,开展大范围高精度跟踪的运动控制策略研究;以插件机为背景,研究高精度快速点到点运动控制问题。本文的工作主要包括如下几个方面: (1)针对完整乒乓球飞行轨迹测量问题,设计了基于四台智能摄像机的多目立体视觉系统,提出了基于摄像机I/O端口的采集同步方法和内存轮换的图像并行采集与处理方法。此外,还设计了多台计算机设备之间的时间同步方法。 (2)针对乒乓球飞行轨迹预测问题,提出了一种乒乓球解析飞行模型,基于梯度下降计算乒乓球的飞行初始速度。利用乒乓球反弹前后轨迹计算乒乓球反弹前后速度,建立了乒乓球在球桌上的反弹模型。针对击球点的选取问题,提出了一种实现机器人最佳运动控制的击球点选择策略。 (3)针对乒乓球机器人球拍规划问题,提出了一种实现定点、定速回球的乒乓球期望出球速度的计算方法。分析了击球点位置和来球速度的预测误差对落点的位置和速度造成的影响。针对机器人打球规划问题,提出了包含迎球阶段、击球阶段和返回阶段的三阶段打球策略,分别设计了基于位置控制和基于速度控制的打球规划方法。 (4)提出了一种基于宏微机器人的宏微协调运动策略,将宏动规划、微动自动调整以及矫正运动结合起来,实现机器人的大范围、高精度焊缝跟踪。 (5)提出了一种基于光纤传感器的插件机机头和PCB板高精度对准方法和插件坐标的逐点纠正和批量纠正方法,实现了插件机高精度快速点到点运动。 最后,对本文所取得的研究成果进行了总结,并指出需要继续开展的工作。
其他摘要Cartesian robots are widely used in the industry applications for the reasons that their structures are very simple, their control systems are easy to realize, and they have strong extensibility. This dissertation does related researches on high speed motion control under visual guidance, large scale tracking with high precision and fast motion control with high accuracy for Cartesian robots. A 5 degree-of-freedoms table tennis robot is used to study the motion planning and high speed motion control methods under visual guidance. A macro-micro robot is adopted to study the motion control strategy of large scale tracking with high precision. An inserter machine is used to investigate fast and high accuracy motion control. The main contributions of this paper are as follows: Firstly, a multi-camera stereovision system based on four smart cameras is developed to measure the complete ball flying trajectories. A Method for image synchronous acquisition based on the smart camera’s I/O ports and a method for image acquisition and processing in parallel based on memory shift are proposed. Besides, a method for time synchronization between different computer devices is presented. Secondly, an analytic ball flying model is proposed to predict the ball flying trajectories. The initial ball velocity is computed based on gradient descent. The table rebound model is developed with the ball veloctities before and after rebounding via the measured ball trajectories is introduced. Then, a striking point choosing strategy to realize the best joint motion control is presented. Thirdly, a paddle motion planning method to compute the desired ball velocity after rebounding on the paddle in order to realize returning ball to the desired position with desired speed is presented. The landing position and speed error caused by the prediction error is analyzed. A robot striking planning strategy including approaching stage, striking stage and returning stage is proposed. The motion planning methods based on position control and speed control are designed, respectively. Fourthly, a coordinated movement method based on a macro-micro robot is presented, which combines macro planning, micro self-adjusting and correct motion. The weld tracking in large scale with high precision is realized using the robot with the proposed method. Fifthly, Methods based on fiber sensor to realize high accuracy alignment between the head of inserter and PCB are developed and point by point correctio...
馆藏号XWLW1577
其他标识符200818014628025
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/6332
专题毕业生_博士学位论文
推荐引用方式
GB/T 7714
杨平. 直角坐标机器人的运动规划及控制策略研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2011.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
CASIA_20081801462802(10500KB) 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[杨平]的文章
百度学术
百度学术中相似的文章
[杨平]的文章
必应学术
必应学术中相似的文章
[杨平]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。