Gaussian hybrid filtering method based on local entropy clustering

Abstract

The invention belongs to the field of signal processing, and discloses a Gaussian filtering method based on a local entropy. The method comprises the steps of: (a), performing a task experiment, and acquiring a signal f; (b), for the signal f, calculating the local entropy E<local>, and dividing the signal into n segments according to a local entropy threshold value E<t>; (c), solving Gaussian model (GMM) parameters (pi,mu,sigma) of the signal f by utilizing expected maximum algorithm iteration; (d), performing Gaussian regression (GMR) on the signal f by utilizing the Gaussian parameters (pi,mu,sigma); (e), observing the effect of the signal obtained by regression, and correcting the E<t>, so that an expected filtering effect is obtained; and (f), writing a real-time filtering program byutilizing the Gaussian model parameters (pi,mu,sigma), and performing real-time filtering. By means of the Gaussian hybrid filtering method based on local entropy clustering disclosed by the invention, real-time online filtering of signals can be realized through a method for combining offline processing data analysis with online filtering; and thus, influence of signal filtering delay on real-time control in real-time control can be solved.

Description

technical field [0001] The invention belongs to the field of signal processing, more specifically, a method of Gaussian mixture filtering based on local entropy is designed. Background technique [0002] The accuracy of the feedback signal is the basis of real-time control. The signals collected by various sensors will inevitably be mixed with random white noise and high-frequency noise. In real-time control, how to quickly obtain effective signals has always been an important research content of real-time control. The traditional Butterworth filtering method adopts the frequency truncation method to filter out the noise of other frequencies except the effective information, but it will introduce a large delay. In real-time control, the delay will lead to the instability of the control system. Therefore, it is of great significance for real-time control based on feedback to filter the original signal with a suitable method. [0003] The patent with the application number 2...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the above defects or improvement needs of the prior art, the present invention provides a Gaussian mixture filtering method based on local entropy, by adopting the principle of local entropy clustering and Gaussian mixture model, thereby solving the problem of real-time control signals based on feedback information Deal with technical issues that delays affect the control system

Images