ANNs to approximate all the inverse kinematic solutions of non-cuspidal manipulators

In this paper a method is proposed for finding the complete number of the inverse kinematic solutions by using of Artificial Neural Networks (ANN). The training data, which are generated using the Halton quasi-random sequence, are clustered into aspects by using the proposed new metric function for...

Full description

Saved in:
Bibliographic Details
Published inIFAC-PapersOnLine Vol. 51; no. 22; pp. 418 - 423
Main Authors Frisyras, Eleftherios K., Moulianitis, Vassilis C., Aspragathos, Nikos A.
Format Journal Article
LanguageEnglish
Published Elsevier Ltd 2018
Subjects
ANN
Determinant of Jacobian
Halton sequence
IKP
Metric function
Multiple Solutions
ANN
Halton sequence
Metric function
Determinant of Jacobian
Multiple Solutions
IKP
Online AccessGet full text

Cover

More Information
Summary:In this paper a method is proposed for finding the complete number of the inverse kinematic solutions by using of Artificial Neural Networks (ANN). The training data, which are generated using the Halton quasi-random sequence, are clustered into aspects by using the proposed new metric function for the joint configuration space. Application of the method in a non-cuspidal manipulator with six (6) Degrees of Freedom (DoF) having sixteen (16) distinct inverse kinematics solutions demonstrates the effectiveness of the proposed approach. Concluding remarks and hinds for future research is closing this paper.
ISSN:2405-8963
2405-8963
DOI:10.1016/j.ifacol.2018.11.586