Frequency-domain identification algorithms for servo systems with friction
Resource
IEEE Transactions on Control Systems Technology 10 (5): 654- 665
Journal
IEEE Transactions on Control Systems Technology
Journal Volume
10
Journal Issue
5
Pages
654-665
Date Issued
2002
Date
2002
Author(s)
Abstract
Mechanical devices usually come with undesirable nonlinearities, such as friction, backlashes, and saturations. Under the assumption of linear systems, the commonly seen identification schemes utilize sinusoidal excitation signals for parameter identification. However, the data needed for identification are unavoidably distorted by the aforementioned nonlinearities and the identification result may not be satisfactory. In the paper, binary test signals are used to perform identification, thus simplifying the behavior of friction. An identification method based on the difference of binary multifrequency excitation signals is proposed. The modified identification algorithm does not suffer from the problem of nonlinear distortions in the signal shape and is able to determine the nonlinear friction such that an accurate servo system model can be derived. A high-precision ball-screw table with asymmetric friction is identified as a test plant for this approach. The results prove that the method can be used very successfully.
Type
journal article
File(s)![Thumbnail Image]()
Loading...
Name
18.pdf
Size
265.01 KB
Format
Adobe PDF
Checksum
(MD5):b5ca60b72361f1939ae685b19b1f30f3