摘要
提出一种以非正交小波为基函数并应用小波多尺度分析的系统脉冲响应辨识方法,该方法以小波级数的形式逼近脉冲响应过程.从理论上证明了经小波尺度变换后系统随机噪声的方差值减小,即噪声得到有效的抑制,从而大大提高了辨识精度.应用实例验证了所得结果的正确性和算法的实用性.
A pulse response identification approach based on orthonormal scaling transformation and non-orthonormal wavelet as basis functions is proposed. Wavelet function expansion is applied to express the pulse response function. The random noises of systems are shown to be restrained effectively through orthonormal scaling transformation, and estimation precision is improved. An application example shows a satisfactory identification result and feasibility of the method.
出处
《控制与决策》
EI
CSCD
北大核心
2004年第10期1190-1193,共4页
Control and Decision
基金
国家自然科学基金资助项目(60374028).
关键词
小波级数
尺度变换
函数逼近
系统辨识
Impulse response
Nonlinear systems
Parameter estimation
Wavelet transforms
