Modelling and control of two-link flexible manipulator

Flexible link manipulators have caught the interest of many researchers due to the limitations of their rigid counterparts. However, Flexible manipulators introduces undesired vibrations which is not easy to control due to its high-non linearity. In order to keep the advantages associated with the l...

Full description

Saved in:
Bibliographic Details
Main Author: Negmeldin, Amr
Format: Thesis
Language:English
Published: 2017
Subjects:
Online Access:http://eprints.utm.my/id/eprint/78480/1/AmrNegmeldinMFKM2017.pdf
http://eprints.utm.my/id/eprint/78480/
http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:109823
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Flexible link manipulators have caught the interest of many researchers due to the limitations of their rigid counterparts. However, Flexible manipulators introduces undesired vibrations which is not easy to control due to its high-non linearity. In order to keep the advantages associated with the lightness and flexibility of the manipulators, accurate modelling of the system and efficient reliable controller have to be developed which is the focus of this study. The two-link flexible manipulator is split into 4 models, the Hub angle and endpoint vibrations of both links of the Two-Link Flexible Manipulator. Input and output data were obtained from an experimental rig. Each model was obtained through system identification techniques within MATLAB simulation environment, namely conventional Recursive Least Square and Cuckoo Search Algorithm. Comparison was made between models developed using the two algorithms and this study shows that Cuckoo Search Algorithm is superior than Recursive Least Square Algorithm based on Mean Square error (MSE). RLS developed models MSE are 5.6321×10−5,0.0018,0.0129 & 0.0078e for hub angle 1, hub angle 2, deflection 1 and deflection 2 respectively. CSA developed models MSE are 2.7164×10−5,1.1546×10−5,6.0404×10−4 & 0.0026 respectively. Correlation tests showed that the hub angle models are biased, while the deflection models are unbiased for both algorithms. Finally, controllers intelligently tuned by Cuckoo search optimization algorithm were introduced to control the hub angle position and the endpoint vibrations. The rise time and maximum overshoot are 0.5 seconds and 0 rad for hub angle 1 and 0.5 seconds and 0.2 rad for hub angle 2. The setting time and maximum overshoot are 2 seconds and 0.01 rad for deflection 1 and 2 seconds and 0.007 rad for deflection 2.