基于抗扰扩展卡尔曼滤波的摇臂伺服控制  

Swing Servo Control Based on Extended Kalman Filter With Disturbance-rejection

作　　者：肖鑫 孙乐 甘辉 XIAO Xin;SUN Le;GAN Hui(School of Automation,Nanjing University of Science&Technology,Nanjing 210094,Jiangsu Province,China)

机构地区：[1]南京理工大学自动化学院,江苏省南京市210094

出　　处：《中国电机工程学报》2023年第20期8102-8113,共12页Proceedings of the CSEE

基　　金：国家自然科学基金项目(51907091);江苏省自然科学基金项目(BK20190467)。

摘　　要：大惯量直驱系统通常需要高增益控制器去实现高响应伺服控制,一般采用位置传感器,经过锁相环配合滤波器对转子位置进行处理,获得转速反馈。常规滤波器的高带宽容易引入转速噪声影响控制稳定性;而低带宽滤波器又限制了控制系统动态响应。针对此现象,该文首先采用基于摇臂运动模型的抗扰扩展卡尔曼滤波器(extended Kalman filter,EKF)观测电机转速,该方法既避免了信号噪声,又便于支持基于模型的运动控制方法获得高控制响应。实际对象运动中模型参数变化较大,例如运动阻尼、电机转矩常数等在不同摇臂位置的数值差别较大,而常规EKF对参数依赖较高。为解决该问题,引入抗扰补偿机制,通过比较EKF位置估计值与实际转子位置测量值偏差,可以获得运动参数变化导致的观测扰动。通过补偿该扰动,可以抑制由于参数变化导致的速度观测误差;再次,采用模型预测控制(model predictive control,MPC)来进一步改善系统动态性能;最后,通过仿真对所提策略进行了研究,并在基于永磁同步电机(permanent magnet synchronous motors,PMSMs)的摇臂伺服实验台上验证了所提方案的有效性。Large inertia direct drive system often needs high gain controller to realize high response servo control,and the rotor position sensor of motor is generally used to process the rotor position through the phase-locked loop and filter to obtain the speed feedback.The high bandwidth of conventional filter is easy to introduce speed noise,which affects the control stability,Also,the low bandwidth filter limits the control response of the system.First,the extended Kalman filter(EKF)based on the rocker motion model is used to observe the motor speed in this paper.This method not only avoids the signal noise,but also is convenient to support the model-based motion control method to obtain high control response.However,the model parameters change greatly in the actual motion,such as the values of motion damping and motor torque constant at different rocker arm positions,which are quite different,so the conventional EKF is highly dependent on the parameters.In order to solve this problem,this paper introduces the anti-interference compensation mechanism.By comparing the deviation between the EKF position estimation and the actual rotor position measurement,the disturbance caused by the change of motion parameters can be obtained.By compensating the disturbance,the velocity observation error caused by parameter change can be suppressed.Furthermore,this paper uses model predictive control(MPC)to further improve the dynamic performance of the system.Finally,this paper studies the proposed strategy through simulation,and the effectiveness of proposed scheme can be proved through the rocker arm servo test-bed based on permanent magnet synchronous motors(PMSM).

关 键 词：伺服驱动 模型预测 永磁同步电机 速度观测 扩展卡尔曼滤波 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]