Noise coefficient computing method used for Allan variance analysis technology

Abstract

The invention discloses a noise coefficient computing method used for the Allan variance analysis technology. The noise coefficient computing method used for the Allan variance analysis technology solves the problems that matrix inversion is singular, noise coefficients have negative values, and the noise coefficients are not optimal, wherein the problems exist in five types of typical random noise coefficient processes in the prior art. The noise coefficient computing method used for the Allan variance analysis technology mainly comprises the following steps that 1) an Allan variance sequence and a corresponding relevant time sequence are preset; 2) a relevant time matrix is computed; 3) a halt condition is set, and linearization is conducted on a target function; 4) initial values of loop iteration related parameters are set; 5) matrix transformation is conducted; 6) sphere transformation is conducted, and an unconstrained optimal increment is computed; 7) a hyperplane nonnegative restriction optimal increment is computed; 8) the noise coefficients are updated; 9) a halt index is computed; 10) whether a halt criterion is met is judged. Random noise coefficients estimated by the noise coefficient computing method can be used for inertial device performance evaluation or accurate modeling of random measurement errors, and has significance for improving navigation accuracy of an inertia matrix integrated navigation system.

Description

technical field [0001] The invention relates to an estimation method of noise intensity coefficient, in particular to a noise coefficient calculation method used in Allan variance analysis technique. Background technique [0002] Allan variance analysis technology refers to a time-domain noise analysis technology that uses the Allan variance of the output data when the instrument is static to analyze the random noise component and its intensity coefficient contained in the output data of the instrument. There are a wide range of applications in performance analysis of components such as gyroscopes and accelerometers. It is generally considered that the output data of the instrument when it is static may contain five typical random noises, namely, quantization noise, white noise, zero bias instability, rate random walk and rate ramp. The estimation of these five kinds of random noise intensity coefficients is an important part of the Allan variance analysis technique. Curren...

AI Technical Summary

Problems solved by technology

[0003] The present invention aims at the problems existing in the estimation of five typical random noise coefficients in the Allan variance analysis technique, and proposes an optimization coefficient estimation method based on non-negative number constraints

Images