Authors:
Donatello Tina
;
Luca Carbonari
and
Massimo Callegari
Affiliation:
Polytechnic University of Marche, Italy
Keyword(s):
Neural Networks, Robot Control, Parallel Kinematics Machines.
Related
Ontology
Subjects/Areas/Topics:
Engineering Applications
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Neural Networks Based Control Systems
;
Robot Design, Development and Control
;
Robotics and Automation
;
Signal Processing, Sensors, Systems Modeling and Control
Abstract:
The paper deals with a neural network control for the gravity compensation of a parallel kinematics robot. The network has been designed in a simulation environment then it has been implemented in robot’s controller by using an FPGA device that is part of an embedded system. After the training phase, several experiments have been performed on the prototype manipulator and the related datasets have been collected and elaborated. In the end, a comparative analysis has shown the improved performance of the neural network controller with respect to the inverse dynamics one, mainly due to the well-known difficulties in deriving explicit models of friction and play in the joints.