Image-based visual servoing (IBVS) is one of the popular VS approaches for robot manipulators since it does not require pose estimation. IBVS has to cope with theoretical problems as well as problems during implementation. The main three of these realization problems are obtaining the inverse of the interaction matrix, defining a suitable constant gain value for the controller, and keeping the features in the field of view, respectively. Although the interaction matrix for IBVS is used with the pseudoinverse, the control law does not work if singularities occur. On the other hand, the constant gain value causes a trade-off between the convergence speed and the end-effector speed. It is also a common problem that the features may leave the field of view during IBVS operation. In this study, it is aimed to implement the intelligent approaches proposed to solve these problems on an industrial type robot manipulator. As the first stage of the implementation, the intelligent approximation units take the place of the inverse of the interaction matrix and the problem of singularity disappears. As a second stage, instead of a fixed gain, a fuzzy logic unit which calculates gain according to the value of error and error derivative in each iteration is used. In the third stage, regions are defined in the image plane, and the features are kept within the field of view with the help of a fuzzy logic unit. In this study, experimental results of all implementations are presented and discussed.
|Thanks||Bu çalışma TÜBİTAK tarafından 3001 kodlu Başlangıç Projeleri kapsamında TÜBİTAK-117E511 numarası ile desteklenmiştir.|
: January 31, 2020
|APA||Yuksel, T . (2020). GTGS Gerçekleme Problemleri İçin Akıllı Çözümlerin Endüstriyel Bir Robot Manipülatöre Uygulanması . Bilecik Şeyh Edebali Üniversitesi Fen Bilimleri Dergisi , 7 (2) , 764-789 . DOI: 10.35193/bseufbd.682875|