Truth value estimation method based on multi-dimensional vector information and device thereof

Abstract

The invention provides a truth value estimation method based on a multi-dimensional orthogonal measurement, comprising using two or more orthogonal measurement units to measure the same measurement object simultaneously in real time to obtain a measurement result judgment function F(R)obtained by the multi-dimensional measurement vector based on the obtained multi-dimensional measurement vector construction, wherein the functions relative to each vector in the multi-dimensional measurement vector are as follows: C(R) is a restraint factor, R is a scanning moving point at a fixed line in 45 degrees, the restraint factor C(R) is used for standardizing the extremum characteristic and divergent characteristic of the judgment function, the extremum is obtained by the judgment function F(R) when the R is the truth value point, the selection thereof is provided with the flexibility, an extremum point of the constructed measurement result judgment function F(R) is used as the truth value point by scanning on the fixed line in 45 degrees, and the value corresponding to the truth value point is the truth value estimation value.

Description

technical field [0001] The present invention relates to the field of communication technology and signal processing, and more specifically, to a method of truth estimation based on multidimensional vector measurement information. The model provides high-precision and high-stability estimates for orthogonal measurements. Background technique [0002] Research and industry have been seeking various ways to improve measurement accuracy. There are roughly two ways to improve measurement accuracy. One way is to improve the sensing accuracy and sensing ability of the measurement device itself; the other is to use the collected data and use the method of statistical signal processing to achieve the goal of improving the measurement accuracy through post-calculation and processing. [0003] For the second method of improving measurement accuracy, at the current research level, for generalized stationary signal measurement, two methods are usually used to improve measurement accura...

AI Technical Summary

Problems solved by technology

However, in various measurement errors, it is extremely difficult to eliminate Gaussian noise with non-zero mean (such as figure 1 ), which becomes the key factor hindering the improvement of accuracy

[0004] Due to the existence of non-zero mean Gaussian noise, the traditional maximum likelihood estimation and moment estimation are used to estimate the true value, which is usually greatly affected by the non-zero mean, so that the obtained estimated value is easy to be far from the true value

In addition, Kalman filtering and Wiener filtering have limited effects on the elimination of this error, so a new truth processing method is needed

Images