Title :
Robot dynamics identification via neural network
Author :
Alexander A. Dyda;Dmitry A. Oskin;Andrey V. Artemiev
Author_Institution :
Department of Automatic and Information Systems, Far Eastern Federal University, 8 Suhanova St., Vladivostok 690950, Russia
Abstract :
Recurrent neural network (RNN) - based approach to identification of underwater robot (UR) is considered and investigated in the paper. It was shown that RNN models can be successfully trained to nonlinear behaviour of a UR. Experiments carried out with data taken from UR dynamics model also confirmed effectiveness and prospective of the approach considered.
Keywords :
"Mathematical model","Robots","Training","Data models","Recurrent neural networks","Testing"
Conference_Titel :
Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS), 2015 IEEE 8th International Conference on
Print_ISBN :
978-1-4673-8359-2
DOI :
10.1109/IDAACS.2015.7341437
