Constrained Linear Regression for Flight Control System Failure Identification

P. Chandler,M. Pachter,M. Mears

DOI: https://doi.org/10.1109/ACC.1993.4176474

1993-06-02

Abstract:Continuous system identification of aircraft stability and control derivatives is proposed for failure identification. Prior information from flight mechanics is incorporated as constraints in linear regression, to improve the speed and accuracy of identification. A linearized fourth - order F-16 model with measurement (sensor) noise and damage to the horizontal tail is simulated. The simulated inflight failure problem demonstrates significant improvements in the short period parameter estimates.