Minimal resource allocating networks for aircraft SFDIA

Presents an online learning approach for the problem of sensor failure detection, identification and accommodation (SFDIA) for aircraft system using neural networks (NNs). The SFDIA scheme exploits the analytical redundancy of the system to provide sensor validation capability to a measurement devic...

Full description

Saved in:
Bibliographic Details
Main Authors: Fravolini, M.L., Campa, G., Napolitano, K., Yongkyu Song
Format: Conference Proceeding
Language:English
Subjects:
Aerospace control
Aerospace electronics
Aerospace engineering
Aircraft propulsion
Fault tolerant systems
Mechanical sensors
Neural networks
Redundancy
Resource management
Sensor systems
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Presents an online learning approach for the problem of sensor failure detection, identification and accommodation (SFDIA) for aircraft system using neural networks (NNs). The SFDIA scheme exploits the analytical redundancy of the system to provide sensor validation capability to a measurement device by employing learning NNs as online nonlinear approximators. In the context of online learning some issues are of critical importance, such as learning speed, number of parameters to be updated, and stability of the learning algorithm. To address these problems a minimal resource allocating network (MRAN) is proposed featuring a fully tuned radial basis functions (RBF). The purpose of the study is to evaluate the performance of this architecture on the NN-SFDIA problem, in terms of robustness and fault detectability for both hard and soft sensor failures. The study has been performed on a detailed nonlinear 6 DOF model of the De Havilland DHC-2 "Beaver" Aircraft.
DOI:10.1109/AIM.2001.936897