DocumentCode
2778814
Title
Improvement of tracking performance of servomechanical system by an accurate four-parameter friction modelling and compensation
Author
Chen, Si-Lu ; Tan, Kok Kiong ; Huang, Sunan
Author_Institution
Nat. Univ. of Singapore, Singapore
fYear
2007
fDate
4-6 Dec. 2007
Firstpage
28
Lastpage
34
Abstract
In this paper, a systematic procedure to identify a comprehensive version of friction model via limit cycle experiments using dual-channel relay feedback apparatus is presented and demonstrated on a 3-D Cartesian robotic system. This friction model, which contains static, Coulomb and viscous friction with Sticbeck effect, can describe the friction behaviour when the servo system operates at both low and high velocity mode. The limit cycle properties of the dual-channel relay feedback system are used for careful tuning of relay gains to run the system on either high or low velocity modes. Sets of explicit formulas are applied for direct computation of model parameters. Only the boundary lubrication velocity is estimated via one-dimensional optimisation. The friction compensation experiment shows the accuracy of the identified model and the effectiveness of proposed method.
Keywords
optimisation; robots; servomechanisms; stiction; 3D Cartesian robotic system; Coulomb friction; Sticbeck effect; boundary lubrication velocity; dual-channel relay feedback; four-parameter friction modelling; friction compensation; limit cycle experiments; one-dimensional optimisation; servomechanical system; static friction; viscous friction; Control systems; Data mining; Feedback; Friction; Limit-cycles; Lubrication; Mathematical model; Relays; Robots; Servomechanisms; Friction modelling; limit cycle; relay feedback;
fLanguage
English
Publisher
ieee
Conference_Titel
Mechatronics and Machine Vision in Practice, 2007. M2VIP 2007. 14th International Conference on
Conference_Location
Xiamen
Print_ISBN
978-1-4244-1358-4
Electronic_ISBN
978-1-4244-1358-4
Type
conf
DOI
10.1109/MMVIP.2007.4430709
Filename
4430709
Link To Document