Inverse kinematic solution of 6-DOF industrial robot using nero-fuzzy technology

Dash, Kshitish K.; Choudhury, B. B.; Senapati, S. K.

doi:10.1504/ijcsyse.2019.10025339

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dereli and Köker [13] introduced an artificial bee colony method for solving the inverse kinematics of a 7-degree-of-freedom robotic arm. Köker et al [14] introduced a three-joint robotic manipulator inverse kinematics solution based on neural networks while Dash et al [15] employed an artificial neural network to tackle a complicated inverse kinematics issue for a 6-DOF Industrial robotic manipulator. They solved the inverse kinematics using a geometric technique.…”

Section: Introductionmentioning

confidence: 99%

Surgical Self- steering Robot using Artificial Neural Network

Mahfouz,

Gamal,

Awad

et al. 2023

Preprint

View full text Add to dashboard Cite

The aim of this work is to design a robot that can be used in surgery and can remove tumors by accurately and safely following contours. This is achieved by the medical imaging set such as DICOM-formatted images from several sources (CT, MRI, and Ultrasound) of these tumors using artificial neural networks. The workspace of a suggested 7 degrees of freedom robot is defined by calculating the position of the end effector for multiple positions of its links using DH tables and feed-forward kinematics. The data obtained are used to train three artificial neural networks to get the inverse kinematics. The suggested neural networks are the Generalized Regression Neural Network, Radial Basis, and Feed Forward. A training course for inverse kinematics of 23328 workspace points of the end effector of the suggested robot is used for training the three neural networks. The most accurately trained one was found to be the Feed-Forward neural network. It shows a root mean square error for those points of 2.08 mm and a training time of 1236 seconds and is to be applied to steer the robot links. The patient medical images are digitized using the image processing toolbox of MATLAB. The data obtained is to be fed to the superior neural network to get the robot links positions needed to move the end effector mounted with a cutter to follow the tumor contour accurately. The absolute error distance between the corresponding points of the contours resulting from testing the trained Feed-Forward NN with 51 points of a contour doesn’t exceed 2.08 mm.

show abstract

Section: Introductionmentioning

confidence: 99%