Design considerations for flight test of a fault inferring nonlinear detection system algorithm for avionics sensorsThis paper summarizes the modifications made to the design of a fault inferring nonlinear detection system (FINDS) algorithm to accommodate flight computer constraints and the resulting impact on the algorithm performance. An overview of the flight data-driven FINDS algorithm is presented. This is followed by a brief analysis of the effects of modifications to the algorithm on program size and execution speed. Significant improvements in estimation performance for the aircraft states and normal operating sensor biases, which have resulted from improved noise design parameters and a new steady-state wind model, are documented. The aircraft state and sensor bias estimation performances of the algorithm's extended Kalman filter are presented as a function of update frequency of the piecewise constant filter gains. The results of a new detection system strategy and failure detection performance, as a function of an update frequency, are also presented.
Document ID
19860062773
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Caglayan, A. K. (Charles River Analytics, Inc. Cambridge, MA, United States)
Godiwala, P. M. (Charles River Analytics, Inc. Cambridge, MA, United States)
Morrell, F. R. (NASA Langley Research Center Hampton, VA, United States)