Recurrent Neural Network-based Fault Detector for Aileron Failures of Aircraft

Nobuyuki Yoshikawa 1 Nacim Belkhir 2 Sinji Suzuki 1
2 TAU - TAckling the Underspecified
LRI - Laboratoire de Recherche en Informatique, UP11 - Université Paris-Sud - Paris 11, Inria Saclay - Ile de France, CNRS - Centre National de la Recherche Scientifique : UMR8623
Abstract : This paper empirically investigate the design of a fault detection mechanism based on Long Short Term Memory (LSTM) neural network. Given an equation based model that approximate the behavior of aircraft ailerons, the fault detector aims at predicting the state of aircraft: the normal state for which no failure are observed, or four different failure states, e.g. a delay changes. This is achieved by collecting a limited amount of command and responses data by varying the parameters of the aileron model, such that a LSTM network is used to predict the state of the aircraft of sequence of the pair commands/responses. In this empirical study we empirically demonstrated LSTM networks can be a promising approach for fault detection, and achieve reasonable performances despite a limited amount of data, in particular avoiding overfitting of the model.
Complete list of metadatas

Cited literature [25 references]  Display  Hide  Download

https://hal.inria.fr/hal-01669540
Contributor : Nacim Belkhir <>
Submitted on : Wednesday, December 20, 2017 - 11:12:32 PM
Last modification on : Thursday, February 7, 2019 - 4:33:06 PM

File

ACSS2017.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01669540, version 1

Citation

Nobuyuki Yoshikawa, Nacim Belkhir, Sinji Suzuki. Recurrent Neural Network-based Fault Detector for Aileron Failures of Aircraft. ASCC 2017 - The 2017 Asian Control Conference , Dec 2017, Gold Coast, Australia. ⟨hal-01669540⟩

Share

Metrics

Record views

490

Files downloads

545