Novel Genetic Algorithm with Parameter Exchanger (GAPE) for Identification of Flexible Manipulator System

• Published in: IET Conference Proceedings
• Main Author: Yatim H.M.; Hadi M.S.; Talib M.H.A.; Mat Daras I.Z.
• Format: Conference paper
• Language: English
• Published: Institution of Engineering and Technology 2022
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85174653928&doi=10.1049%2ficp.2022.2250&partnerID=40&md5=402efb207e722a49a6de8164ab0d2c07

As technology advances, more applications require lighter, more flexible and less bulky structures. Thus, flexible manipulator has been favoured in robotics and mechanical systems. This paper presents the modeling of single-link flexible manipulator system using an intelligent optimizer, Genetic Algorithm with Parameter Exchanger (GAPE). In this highly complex system where unwanted vibrations occur during operation, GAPE was developed to accurately model this system. A single-link flexible manipulator experimental setup was initially developed. Hub angle and end-point acceleration experimental data was collected and fed to the system identification method. After that, identification was carried out by applying the proposed novel GAPE through a linear auto regressive with exogenous (ARX) model structure as compared to a standard Genetic Algorithm (GA). Algorithms was then validated based on the reducing value of mean-squared error (MSE) and correlations test. From the results, the advantages of the 'parameter exchanger' was confirmed where GAPE achieved the smallest MSE value of 2.9567 × 105 and 6.4614 × 106 for end-point acceleration and hub angle modelling, respectively. Good correlation values was also achieved by both models which indicating within 95% confidence interval. In the future, validated model will serve as a platform for implementation of active vibration control of flexible manipulator systems. © 2022 IET Conference Proceedings. All rights reserved.