摘要
为深入认识摩擦对航空相机扫描镜系统频率响应特性的影响,获得更准确描述该系统真实动态的模型,使用随机相位多正弦信号测量了扫描镜系统的频率响应特性并进行了线性近似参数模型辨识。首先,介绍扫描镜系统辨识的实验平台与激励信号选择。然后,使用奇-奇频率随机相位多正弦信号分别测量扫描镜系统在非激励频率处和激励频率处的输出对输入信号幅值的依赖,从而定量评估摩擦非线性的影响。最后,基于信号采样均值及噪声采样方差、协方差估计辨识了扫描镜系统线性近似参数模型。实验结果表明,扫描镜系统的摩擦非线性主要出现在奇频率处,高于噪声10dB;系统的频率响应特性依输入信号幅值不同而各异,在低于20rad/s频率区该差别尤为显著。由于摩擦非线性影响,扫描镜系统需要使用3阶模型描述;与正弦扫描方法相比,基于多正弦信号激励获得的参数模型可更好地描述扫描镜系统真实动态特性。得到的结果为控制器的设计奠定了基础。
To know the effect of friction on the frequency response of Scanning Mirror System（SMS） in an aerial camera and to provide a model for describing the real dynamic of the SMS more exactly,the random phase multi-sine exciting signal was used to detect the feature of frequency response of the SMS and to identify a linear approximate parameter model.Firstly,the experiment setup for SMS identification was introduced and the exiting signal was selected.Then,the odd-odd frequency random phase multi-sine signal was used to measure the dependence of output of system at the non-exited frequency and exited frequency on the input signal,and to evaluate the nonlinear effect of friction quantitatively.Finally,an estimation method based on the sample mean value of signal and sample noise（co-）variances was used to identify the linear approximate parameter model of SMS.The experimental results indicate that the friction nonlinearity in the SMS is mainly located at the odd frequency,10 dB above the noise.The frequency response of SMS varies under the different amplitudes of input signal,and the difference is obvious especially below 20 rad/s.Because of the effect of friction,the SMS should be described by a 3-order parameter model.Compared with sine-swept method,the parameter model based on multi-sine signal is better in describing the real dynamic feature of the SMS.Obtained results can provide the base for designing of controllers.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2013年第7期1771-1779,共9页
Optics and Precision Engineering
基金
国家863高技术研究发展计划资助项目(No.2010AA010102)
关键词
航空相机
扫描镜
摩擦
随机相位多正弦
线性近似模型
系统辨识
aerial camera
scanning mirror friction
random phase multi-sine
linear approximation model
system identification
