A Data-Driven Controller for Model Unknown Systems

Abstract

The invention provides a data-driven controller for a model-unknown system, which takes a first-order nonlinear system as the controlled object, the input end of the controller is connected with the reference signal, and the output end of the controller is connected with the input end of the controlled system, including Extended state observer module, parameter identification module, data-driven module and linear control module. The data-driven total disturbance and control coefficient estimation method proposed in the present invention can not only estimate the control parameters, but also realize the unified estimation of internal and external disturbances. control precision.

Description

A Data-Driven Controller for Model Unknown Systems technical field The present invention relates to uncertain nonlinear system control field, relate in particular to unknown first-order time-varying uncertainty of control parameters Determining the structure of a nonlinear system controller. Background technique [0002] Many practical control systems, such as motor systems, robotic systems, aircraft systems, ship autopilot systems, etc., are It has the characteristics of time-varying nonlinearity, parameter uncertainty, etc., and there are external disturbances, which bring difficulties to the system control. with challenges. Therefore, for the model unknown system, an estimation method of model uncertainty, external disturbance and control coefficient is proposed to realize The precise control that does not depend on model parameters has important research significance and theoretical value. The adaptive parameter identification method is applied in the design of ...

AI Technical Summary

Problems solved by technology

[0007] According to the above proposal, when the existing control method is applied to the first-order uncertain nonlinear motor system, it cannot deal with the nonlinear uncertainty of the system and the real-time external disturbance, it needs to assume that the control coefficient is known, and the convergence of the estimation error depends on continuous excitation conditions and other technologies problem, and provide a data-driven controller structure for model-unknown systems, which can not only estimate the time-varying uncertainties and external disturbances in nonlinear motor systems, but also estimate the control coefficients of motor systems online

Images