Hardware implementation of a neural network based robust sensor fault accommodation system in flight control system

Abstract

This paper deals with hardware implementation of the robust sensor fault detection and accommodation system of F8 aircraft control system. Both detection and accommodation of normal acceleration sensor is done using two different models of neural network. The simulation model of the novel method of fault accommodation is tested for timing constraints in the embedded system target of ARM processor. Longitudinal plant dynamics of F8 aircraft model and C-Star controller is used for validation of neural network based sensor fault detection and accommodation (NNSFA) system. The robust sensor fault accommodation system is implemented on ARM target of Atmel SAM3X8E ARM Cortex-M3 CPU. This board is inbuilt with ADC and DAC ports to interface pilot input command and the display of the cockpit.