Title :
Calculation of robot parameters based on neural nets
Author :
Lesewed, Ali ; Kurek, Jerzy
Author_Institution :
Inst. of Autom. Control & Robotics, Warsaw Univ. of Technol., Warszawa, Poland
Abstract :
The paper describes applications of recurrent neural network and back propagation learning method for calculation of mathematical model for PUMA 560 robot. The model is based on the Lagrange-Euler formulation and described by a set of nonlinear differential and algebraic equations. A numerical example has shown the comparison of neural model and robot manipulator.
Keywords :
backpropagation; linear algebra; manipulator dynamics; nonlinear differential equations; nonlinear systems; recurrent neural nets; uncertain systems; Lagrange-Euler formulation; PUMA 560 robot; algebraic equations; back propagation learning; mathematical model; neural model; neural nets; nonlinear differential equations; recurrent neural network; robot dynamics; robot manipulator; robot parameters; Differential equations; Friction; Lagrangian functions; Manipulator dynamics; Neural networks; Nonlinear equations; Recurrent neural networks; Robot control; Symmetric matrices; Uncertainty; Puma 560 robot; dynamic model; neural nets;
Conference_Titel :
Robot Motion and Control, 2005. RoMoCo '05. Proceedings of the Fifth International Workshop on
Print_ISBN :
83-7143-266-6
DOI :
10.1109/ROMOCO.2005.201411
