|
AUTHOR(S): Satish Gummadi Saalma
|
TITLE Neural Network Based State Observer with Unknown Terms for Actuator Fault Approximation |
 PDF |
ABSTRACT In this article, neural network based observer with unknown terms has been implemented for actuator fault approximation in linear systems. All states in a system are not measurable or difficult to measure. In such cases, state observers are used to estimate the states. In this paper, state observer problem with unknown terms is considered. Neural networks are used to approximate the unknown terms in the model. The neural network is trained using back propagation algorithm.The proposed observer is tested on DC motor with various actuator faults such as abrupt, incipient and sinusoidal faults. It is observed that the results obtained for these faults are validated the satisfactory performance of the observer. |
KEYWORDS State observer, Neural networks, Actuator faults |
|
Cite this paper Satish Gummadi Saalma. (2020) Neural Network Based State Observer with Unknown Terms for Actuator Fault Approximation. International Journal of Circuits and Electronics, 5, 35-43 |
|
