A semiparametric estimation method for aircraft trajectory

DOI:	10.11809/bqzbgcxb2024.02.026
Keywords:	aircraft trajectory; systematic error; semiparametric estimation; nonlinear least square; wavelet analysis
Abstract:	A new estimation method of aircraft trajectory based on semiparametric modeling was proposed in this paper. This method improved the estimation accuracy of trajectory by mitigating the measurement systematic errors which could not be parameterized. The parametric model of aircraft trajectory was constructed by using the B spline function with non uniform knots, and this model was embedded into the measurement model. The systematic errors which could not be parameterized were expressed as non parameters in the measurement model. Then, a semiparametric regression model was established containing two types of unknown variables to be estimated: the trajectory parameters and the non parameters of systematic errors. Trajectory parameter estimation and systematic error separation were operated alternately and iteratively to solve this semiparametric regression model. The trajectory parameters were estimated by using nonlinear least square method, and the systematic errors were separated from residual data by wavelet analysis. Both theoretical analysis and simulation results showed that the trajectory estimation of semiparametric estimation method was more accurate than that of the parametric estimation method which ignored the systematic errors.
Published:	2024-02-28
PDF