Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Nonlinear system identification method based on tensor product network B spline

A technology of nonlinear systems and identification methods, applied in special data processing applications, design optimization/simulation, etc., to achieve the effect of ensuring monotonic convergence and numerical stability

Pending Publication Date: 2021-12-10
NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The method is based on tensor network B-spline (Tensor Network B-splineTNBS) to describe the nonlinear dynamic system, and uses the iterative method of alternating linear scheme to directly estimate the low-rank tensor network used to approximate the high-dimensional B-spline To solve multivariate B-splines will lead to the curse of dimensionality

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Nonlinear system identification method based on tensor product network B spline
  • Nonlinear system identification method based on tensor product network B spline
  • Nonlinear system identification method based on tensor product network B spline

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0083] The present invention will be further described below in conjunction with the accompanying drawings.

[0084] Aiming at the characteristics of hysteresis and noise in nonlinear systems, the present invention proposes a nonlinear system identification method based on tensor product network B-splines, which can identify nonlinear dynamic systems with a large number of hysteresis and inputs and ensure monotone convergence , and the numerical stability is ensured by the orthogonality of the TT kernel. This method first uses the nonlinear autoregressive exogenous NARX model to describe the nonlinear dynamic system; then identifies the NARX model based on the Tensor Network B-spline TNBS, and uses the An iterative approach in alternating linear schemes that directly estimate low-rank tensor networks for approximating high-dimensional B-splines, thus eliminating the need to explicitly construct exponentially weighted tensors; finally numerical stability is ensured by the ortho...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a nonlinear system identification method based on a tensor product B-spline, which comprises the following steps: firstly describing a non-linear dynamic system by adopting a non-linear autoregression model NARX, then identifying the NARX model based on a tensor product network B-spline, and identifying the non-linear dynamic system by adopting an iterative method based on an alternating linear format; directly estimating a low-rank tensor network used for approaching a high-dimensional B spline to eliminate the need for explicitly constructing an index multiple weight tensor; finally, ensuring the numerical stability by orthogonality of the TT kernel so that the robustness to noise is achieved, monotonic convergence is ensured, and a robust identification result of a nonlinear system is obtained. The efficiency and accuracy of the identification method are proved through numerical experiments on a nonlinear system.

Description

technical field [0001] The invention relates to the technical field of system identification, and mainly relates to a nonlinear system identification method based on tensor product network B-splines. Background technique [0002] Estimating predictive mathematical models from data plays an important role in many areas of engineering and science, especially when first-principles based models are not available or are too complex. This mathematical model form should be simple and easy to understand in order to further understand the dynamic characteristics of the unknown system. Due to its simple structure and linear properties, linear models are widely used in system identification. However, for practical systems, there are usually response lags and noise interference, which makes simple linear models unable to describe the complex dynamic behavior of nonlinear systems well. Therefore, it is of great significance and application prospect to study the nonlinear model with str...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F30/20
CPCG06F30/20
Inventor 曹瑞陆宇平姚德清甄子洋尹楚吴震
Owner NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products