Backlash operator and neural network-based adaptive filter

Abstract

The invention relates to a Backlash operator and neural network-based adaptive filter and a method for modeling a creepage and hysteresis system and relates to the technical field of nonlinear hysteresis and creepage system modeling. The Backlash operator and neural network-based adaptive filter comprises a creepage module, a plurality of Backlash operator modules with the same width, a neural network module and an error calculation module. The design method for the adaptive filter is simple and practicable, delay items are avoided, the output of the filter is only related to the current input, and the hysteresis linearity can be reflected simultaneously. Meanwhile, the output of a Backlash operator serial structure is connected with the neural network, the weight can be adjusted in real time, massive sample data is not needed to be prepared in advance, and the modeling accuracy can be greatly improved. The method is suitable for modeling a piezoelectric ceramic actuator, a magnetostrictive actuator, an excited motor and other systems with hysteresis and creepage nonlinear characteristics.

Description

technical field [0001] The invention relates to an adaptive filter based on a Backlash operator and a neural network and a modeling method thereof for a creep hysteresis system, belonging to the technical field of nonlinear hysteresis creep system modeling. Background technique [0002] In recent years, sensors and actuators made of smart materials have been widely used in precision machining and precise positioning systems. However, the nonlinear characteristics of smart materials, such as hysteresis and creep, reduce the repeatability of these systems, slow down the transient response speed, reduce the control accuracy and even make the closed-loop system unstable, and increase the difficulty of controller design. These problems make it difficult for both classical control theory and modern control theory to effectively control them, so it is necessary to propose specific modeling and control methods to solve problems in this field. To compensate creep hysteresis nonlinea...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem that the existing Backlash operator superposition modeling method is complex and the time delay line transverse adaptive filter has insufficient precision when modeling the hysteresis nonlinearity, the present invention proposes a method based on the structure of the transverse adaptive filter. Adaptive Filter of Backlash Operator and Neural Network and Its Modeling Method for Hysteresis-Creep System

Images