摘要
在机器人打磨中,机器人末端六维力/力矩传感器受打磨轮接触振动等影响,力测量信号淹没于噪声中,对该信号实时有效地提取是控制力的关键。从力测量信号快速傅里叶变换后的特征得知,其噪声是由高斯白噪声、振动噪声及有色噪声叠加所构成的复杂噪声,由此提出了基于串联结构的双卡尔曼滤波(double Kalman filter,DKF)算法。包括:第一个滤波采用卡尔曼滤波(Kalman filter,KF)算法,主要滤除高斯白噪声和高频振动谱峰群;针对未能滤除的高斯白噪声和低频振动噪声叠加形成有色噪声的特点,进行了分析,引入单个参数,以指数加权方式设计了一个逐渐时变噪声方差,描述有色噪声特性,以此改进KF作为第二个滤波算法;两个滤波器以串联方式构成双卡尔曼滤波。以眼镜框与打磨轮接触打磨过程为例,试验结果表明,DKF算法比KF算法更加有效地滤除力测量信号噪声,并具有计算复杂度低、实用性强特点。
The consequence that the robot end’s six-dimensional force/torque sensor affected by the contact vibration of the grinding wheel causes the force measurement signal to be submerged by noise in robot grinding,so that real-time and effective extraction of the signal is the key to control force.From the characteristics of the force measurement signal after fast Fourier transform,its noise is a complex noise formed by the superposition of Gaussian white noise,vibration noise,and colored noise.A double Kalman filter(DKF)algorithm based on a series structure was proposed,including:1)the Gaussian white noise and high-frequency vibration peak groups were filtered by the first Kalman filter(KF)algorithm;2)to analyze the characteristics of colored noise formed by the superimposition of unfiltered Gaussian white noise and low-frequency vibration noise,a single parameter was introduced to design the time-varying variance to describe them by exponential weighting,so that KF was improved as the second filtering algorithm;3)two filters formed the double Kalman filter in series.Taking the contact grinding process between eyeglass frame and grinding wheel as an example,the experimental results show that the proposed algorithm is more effective than the traditional KF algorithm in suppressing noise in force measurement signal,and has low computational complexity and strong practicability.
作者
党选举
林智武
DANG Xuanju;LIN Zhiwu(School of Electronic Engineering and Automation,Guilin University of Electronic Technology,Guilin 541004,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2023年第12期220-226,共7页
Journal of Vibration and Shock
基金
国家自然科学基金(61863008,62263004)。
