MEMS gyroscope noise estimation and filtering method

Abstract

The invention provides an MEMS gyroscope noise estimation and filtering method aiming at the problems that an MEMS gyroscope is low in measurement precision, and random noise of the MEMS gyroscope hasuncertainty and nonlinearity. firstly, acquiring a signal with noise and preseeting and initializing a system characteristic function;then calculating a Sigma point, and constructing a statistical property coefficient at the same time; measuring and updating a one-step state prediction mean value, an error covariance matrix, a prediction mean value and a covariance matrix; observing and updatingthe minimum variance; carrying out residual variance updating by utilizing a maximum expectation algorithm and a maximum posteriori estimation criterion; and outputting the updated signal data. According to the method, a suboptimal unbiased MAP noise statistical estimation model is constructed according to a maximum posteriori estimation principle; And on the basis, a maximum expectation algorithmis introduced to convert a noise estimation problem into a mathematical expectation maximization problem, dynamic adjustment of an observation noise variance is realized, and finally estimation and filtering processing of a gyroscope random drift error are realized.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a MEMS gyroscope noise estimation and filtering method. Background technique [0002] In recent years, with the rapid development of Micro Electro Mechanical System (MEMS, Micro Electro Mechanical System), MEMS-based gyroscopes are playing an increasingly important role in inertial navigation. However, the MEMS-based gyroscope has low precision due to its manufacturing process, use environment and other factors, which limits its application. The study found that deterministic error and random noise are two important factors affecting the accuracy of MEMS gyroscope. Deterministic errors can generally be eliminated by algebraic compensation calculation methods, while random noise is an important factor affecting the accuracy of MEMS gyroscopes, which cannot be processed by simple methods, and it determines the bias stability of gyroscopes. Therefore, how to perform effe...

AI Technical Summary

Problems solved by technology

[0003] At present, the commonly used method is to eliminate the random noise of the gyroscope through Kalman filtering, but when the noise has serious nonlinearity, the error is large; some literatures propose an adaptive neural network filtering method, although it has the advantages of online learning, but The amount of calculation is too complicated to realize; there is also the use of UKF (Unscented Kalman Filter, UKF) to achieve nonlinear filtering processing, but UKF filtering performance depends on the prior statistical information of system noise, and inaccurate statistics will cause its filtering divergence; for The shortcomings of UKF, a method of using UKF residual sequence and innovation sequence to estimate the noise characteristics of the system online and improve the adaptive ability of UKF is proposed, but the covariance matching method has a steady-state estimation error, which limits the filtering accuracy

Images