【摘要】 建立了基于小波去噪的增强多尺度自回归模型。增强多尺度状态可以将直接基于小波变换的多尺度自回归模型的网状结构简化为二叉树结构。基于小波去噪的多尺度模型具有非参数化特性 ,该模型适合于系统特性不知道的分布式多分辨率多传感器进行建模。最后 ,应用该算法于高精度划线切割机器人系统中多传感器的建模 ,实现了型钢划线切割过程中型钢边缘的检测。实验结果和数值仿真表明 ,多分辨率多传感器的数据融合可以消除噪声的干扰 ,提高检测系统的测量精度。更多还原

【Abstract】 A class of augment multiscale autoregression model based on wavelet denoise was introduced. The augment multiscale model can simplify the net structure based on wavelet as quadratic tree structure. Multiscale model based on wavelet denoising possesses non-parametric regression. So the model can be applied to solve the problem of distributed multiresolution data fusion. Then the model was applied to model multisensor in high precision marking-cutting robot system and finished the measurement of profiled bars’ edge in the marking and cutting process. The experimental and numerical analysis results show that multiresolution multisensor data fusion can denoise and enhance the precision of the measuring system.更多还原