引用本文:杨 辉,郝丽娜,陈洋,薛帮灿.针对气动肌肉仿生肘关节抖振现象的Kalman-PID控制[J].控制理论与应用,2017,34(4):477~482.[点击复制]
YANG Hui,HAO Li-na,Chen Yang,Xue Bang-can.Kalman-PID control for chattering phenomena of bionic elbow joint actuated by pneumatic artificial muscles[J].Control Theory and Technology,2017,34(4):477~482.[点击复制]
针对气动肌肉仿生肘关节抖振现象的Kalman-PID控制
Kalman-PID control for chattering phenomena of bionic elbow joint actuated by pneumatic artificial muscles
摘要点击 2501  全文点击 2020  投稿时间：2016-07-21  修订日期：2017-01-18
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2017.60535
  2017,34(4):477-482
中文关键词  气动人工肌肉(PAM)  PID控制参数  卡尔曼PID(KPID)  仿生肘关节
英文关键词  pneumatic artificial muscles (PAM)  PID control parameters  Kalman-PID (KPID)  bionic elbow joint
基金项目  国家自然科学基金;国家高新技术研究发展计划
作者单位邮编
杨 辉 东北大学机械工程与自动化学院 110819
郝丽娜* 东北大学机械工程与自动化学院 110819
陈洋 东北大学机械工程与自动化学院 
薛帮灿 东北大学机械工程与自动化学院 
中文摘要
      本文为了抑制气动人工肌肉(PAM)抖振现象, 首先利用PID控制律近似代替其数学模型, 求得其离散状态方程并代入到离散卡尔曼递推公式中, 进而提出基于PID控制参数的卡尔曼PID(Kalman-PID, KPID)控制算法. 为了 验证算法的有效性, 以3自由度PAM仿生肘关节为控制对象, 分别利用PID及KPID控制器对其进行位姿控制. 由实验结果可知, 该算法相较PID控制器拥有更高的控制精度, 提升了系统的鲁棒性, 能够有效抑制由系统过程误差及测量误差所引起的PAM抖振现象, 从而使仿生肘关节运动更加平稳. 此外, 不同于传统卡尔曼滤波算法与控制算法相结合的方式, KPID控制算法无需事先知道被控对象精准的数学模型及噪音特性的先验知识, 从而避免复杂的数学建模过程, 扩大了卡尔曼滤波算法的应用范围.
英文摘要
      For restraining the chattering phenomena of pneumatic artificial muscles (PAMs) caused by process errors and measurement errors, the mathematical model of a PAM was replaced by PID control law whose discrete state equations was solved and taken into the discrete Kalman recursive formulas. Then, the Kalman-PID (KPID) controller based on the PID control parameters was proposed. In order to verify effectiveness of KPID controller, a 3-DOF bionic elbow joint actuated by three pneumatic artificial muscles was treated as the controlled object and controlled via PID controller and KPID controller, respectively. From the results, KPID controller has higher control precision and better robustness than PID controller. KPID controller can restrain the chattering phenomena effectively, which makes the bionic elbow joint rotate smoothly. Moreover, KPID controller does not need the accurate mathematical model of controlled objects and priori knowledge of the noise characteristics, which avoids complicated modeling process and expands the application range of Kalman filtering algorithm.