相位优化的压电纳米定位系统GADRC

Phase Optimized Piezoelectric Nano Positioning System GADRC

压电驱动器的迟滞严重影响压电驱动的定位系统的性能。广义自抗扰控制以线性模型为基准模型,只需估计、补偿非线性和未精确建模的部分模型,进而将实际系统转换为线性系统。为进一步降低迟滞对系统性能的影响,利用微分信号的相位超前性,以总扰动观测值的微分信号优化总扰动估计的相位,提出相位优化的广义自抗扰控制。从参考模型的角度解释扩张状态观测器,获得广义自抗扰控制的参数整定规律。纳米定位实验结果表明,相位优化的广义自抗扰控制的动态响应、抑制干扰和迟滞的能力都优于广义自抗扰控制。

Considering the original ORB algorithm’s the higher detection density of feature points, lower noise immunity ability and poorer adaptation performance, an improved feature extraction algorithm with uniform anti-noise adaptive performance is proposed. Firstly, a Fast Adaptive Image Pyramid(FAIP) method was used to construct the multi-scale layer by Gaussian filtering, difference, normalization and down-sampling of the original input image.Moreover, based on the image layer scale, the number of feature points to be extracted was constrained step by step,so that it can improve the efficiency of the algorithm under the condition of better scale adaptation performance. Secondly, in order to enhance the anti-noise ability and the rotation adaptive performance of features, a Three-Dimensional Descriptor Gray Centroid(3D-DGC) method was proposed, in which horizontal and vertical gradients were added to form a 3D descriptor vector. Finally, a Quadtree Feature Management(QtFM) algorithm was used to select the feature points with Harris maximum in the quad window of each layer to optimize the uniformity of feature points.Experimental results show that, compared with the original ORB algorithm and other improved algorithms, the feature points extracted by the proposed algorithm are more robust.