On Robust State Estimation for Practical Problems in Flight Path Reconstruction

A state estimation method was developed for general flight path reconstruction applications in which the reliability of airdata measurements may vary greatly over time. The method was developed through the application of robust cost functions to the maximum a posteriori estimation philosophy to achieve a trajectory estimation scheme that accurately treats various forms of process noise and time-varying measurement error characteristics that occur in both aircraft system identification and flight path reconstruction applications. It was shown that the estimator is realized as an adaptive maximum a posteriori estimator in which the trajectory and time-varying measurement error covariance matrix are estimated jointly. The efficacy of the scheme was demonstrated through a case study considering its application to simulation data and comparing the performance of implementations realized using a variety of robust cost functions, including a cost function proposed by the authors. A major finding of this work was that implementation based on the authors’ proposed cost function resulted in superior performance. The efficacy of the estimator was further demonstrated though a case study considering its application to real flight-test data. Findings from this case study were consistent with findings from the case study based on simulation data.