System Identification Device for Optimal Dimension Determination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional system identification devices face challenges in determining an optimal system dimension due to unclear boundaries between significant and insignificant singular values in block Hankel matrices, often requiring trial and error, especially when singular values decrease monotonically.
Innovation Solution
A system identification device that receives an impulse response and a designated search range, utilizing a direct feedthrough term identification unit, block Hankel matrix generator, singular value decomposition unit, system dimension determination unit, and system matrix identification unit to determine and output the system dimension based on system characteristics and actual system characteristics, eliminating the need for trial and error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the first system dimension determination method (operator determination based on singular value significance) is used, then the system dimension can be determined by operator judgment, but the determination becomes subjective and may not always yield the optimum system dimension, requiring trial and error
Solution Approach 1:
The system automatically determines the system dimension by comparing calculated system characteristics with actual system characteristics, eliminating the need for operator intervention and trial-and-error processes. The determination is performed autonomously based on objective criteria.
Solution Approach 2:
The system uses feedback from the comparison between calculated system characteristics (from singular value decomposition) and actual system characteristics to automatically adjust and determine the optimal system dimension, creating a closed-loop determination process.
2Extent of automation
If the second system dimension determination method (evaluation function based on singular value rate of change) is used, then automatic determination is achieved, but the determined system dimension changes depending on the evaluation function scheme, still requiring trial and error
Solution Approach 1:
The system performs automatic self-determination of system dimension by autonomously comparing calculated characteristics with actual characteristics, without requiring external evaluation function schemes or operator intervention.
Solution Approach 2:
The patent replaces the mechanical approach of using evaluation functions with a mathematical comparison approach, substituting the need for arbitrary evaluation schemes with objective characteristic comparison.
3Loss of information
If singular value decomposition is applied to block Hankel matrix from actual impulse response, then system matrices can be calculated, but the singular values gradually and monotonously decrease making the boundary between significant and insignificant values unclear
Solution Approach 1:
The system uses feedback from comparing calculated system characteristics with actual system characteristics to objectively determine which singular values are significant, replacing subjective boundary judgment with objective verification.
Solution Approach 2:
The patent replaces the mechanical judgment of singular value significance boundaries with a mathematical comparison system that objectively determines significance based on characteristic matching.
Data Source
AI summary
A system identification device includes: a direct feedthrough term identification unit that receives an impulse response of a dynamic system; a block Hankel matrix generation unit; a singular value decomposition unit that, by singular value decomposition of the block Hankel matrix, outputs a first orthogonal matrix, a second orthogonal matrix, and a singular value; a system dimension determination unit that, on the basis of the first orthogonal matrix, second orthogonal matrix, singular value, and search range, identifies a system matrix other than a direct feedthrough term, and from a comparison of the actual system characteristics and system characteristics calculated on the basis of the system matrix and direct feedthrough term, determines the system dimension; and a system matrix identification unit that identifies a system matrix other than the direct feedthrough term on the basis of the first orthogonal matrix, second orthogonal matrix, singular value, and system dimension.


