一种提取系统脉冲响应函数的高效的去噪算法  

An efficient de-noising algorithm for extracting the impulse response function of system

作　　者：陈恩伟[1] 刘正士[1] 陆益民[1] 王勇[1] 

机构地区：[1]合肥工业大学噪声振动工程研究所,安徽合肥230009

出　　处：《振动工程学报》2005年第4期475-479,共5页Journal of Vibration Engineering

基　　金：国家自然科学基金资助(60175027);国家863资助(2001AA423300)

摘　　要：针对锤击激励方式不适用于工业机器人操作臂动态性能测试的情况,根据Duharm a l积分理论,利用系统输入矩阵的上三角形的特殊结构,并结合时域平均技术,导出了具有抗噪性能的高效的求取系统脉冲响应函数的平均递推算法。研究了算法对适合于机器人操作臂测试的阶跃激励方式的适用性,对算法的误差进行了分析。在二自由度动力学模型上进行仿真,与传统的时域解矩阵方法比较,平均递推算法在效率及降噪效果上具有明显的优越性。悬臂梁实验表明,平均递推算法能够获得与锤击法比较一致的结果。In the case of inapplicability of hammering excitation to the arm of industrial robot for extracting its dynamic characteristics, according to the Duharmal integral theory and applying the specific upper triangle construction of input matrix of system as well as the technique of time average, an efficient de-noising average recursive algorithm is presented in this paper. Applicability of this method for step excitation, which is suitable for extracting dynamic characteristics of arm of robot, is studied. The error of the algorithm is analyzed and some conclusions are arrived. Results of simulation in 2-DOF dynamic model show that, in contrast with traditional time-domain method, the average recursive algorithm has a better performance on de-noising and is more efficient. Experiment on a cantilever beam shows that the average recursive algorithm can obtain resuits similar to those by hammering excitation method in extracting impulse response function of the systems.

关 键 词：信号处理 脉冲响应函数 时域平均 阶跃激励 递推算法 

分 类 号：TN911.7[电子电信—通信与信息系统]