Discusses the application of recurrent neural networks for identification and control of helicopter vibrations. A class of recurrent networks called memory neuron networks are used for plant identification and control. These networks are obtained by adding trainable temporal elements to feedforward networks. This makes the network output history sensitive and gives them the capability to identify and control systems whose order is unknown or systems with unknown delay. A representative analytical model with higher harmonic pitch angles for minimizing hub shear forces is used for simulation. The effectiveness of the controller in minimizing the force level at varying and constant forward speed are studied. The ability of the controller to cope with changes in system and environment parameters is also considered.
Neural network controller for minimizing hub shear forces in helicopter
1998
5 Seiten, 14 Quellen
Aufsatz (Konferenz)
Englisch
Neural network based feedback error controller for helicopter
Emerald Group Publishing | 2011
|Neural network based feedback error controller for helicopter
Online Contents | 2011
|Vertical-plane pendulum absorbers for minimizing helicopter vibratory loads
Tema Archiv | 1974
|Combined Helicopter Flight Controller
NTIS | 1975
|