Naslov
The Improving of Neural Network Capabilities in On-Line Identification and Tracking Control of Ship

Autori
Velagić, Jasmin ; Vukić, Zoran

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the 10th IEEE International Conference on Methods and Models in Automation and Robotics : MMAR 2004 ; Vol. 1 / - Szczecin : Wydawnictwo Uczelniane Politechniki Szczecinskiej, 2004, 195-200

ISBN
83-88764-09-8

Skup
10th IEEE International Conference on Methods and Models in Automation and Robotics(MMAR) - Special Invited Session "Advanced Ship Control Systems"

Mjesto i datum
Międzyzdroje, Poljska, 30.08.2004. - 02.09.2004

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Neural networks; adaptive learning rate; identification; tracking control

Sažetak
The paper proposes a computationally efficient artificial neural network model for on-line system identification of nonlinear systems under the fuzzy closed-loop control system. The proposed backprogagation (BP) algorithm with adaptive learning rate (BPLAR) was tested for both off-line and on-line identification, comparing with traditional backpropagation learning algorithm on nonlinear ship model. The disadvantages of conventional BP algorithm are slower convergence and longer training times. The learning rate is adjusted at each iteration for the on-line weight and bias adaptation to reduce the training time. Simulation results indicate a superior convergence speed and better control performance in the case of adaptive BP method

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika

POVEZANOST RADA