This paper proposes several improved CMOS analog circuits for neuro-fuzzy network,including Gaussian-like membership function circuit,minimization circuit,and a centroid algorithm defuzzier circuit without using division.

This method transformed star location data with trail into approximately ideal star location data according to characteristic of star location with trail,and then used the centroid algorithm to calculate star location.

Treatment flow for video-image is presented,using AVI function to discompose frames,applying one-dimension maximum entropy algorithm in segmentation of image,employing barycenter algorithm to calculate object s movement track.

This method introduces a subpixel technique and adopts the barycenter algorithm to compute the barycenter of characteristic speckle, which can avoid the complex process of computing the correlation coefficient in digital speckle method; experimental distortion curve for the measured object can be obtained through the related formulas of displacement.

It brought forward the route arithmetic and the weight arithmetic which based on the model,and then the answer of logistic capacity of plant can get from the model.

In this paper we give vector algorithm according to the character of a set of nosinusoidal family orthogonal matrix, these algorithms are simple in design, but are evident in effect.