Research on the Weighted Regression Method of Artillery Trajectory Fitting Computation

Abstract

Abstract In order to improve the precision of the heteroscedastic artillery firing table, a fitting computation method for artillery firing table compilation without the assumption of homologous variance is proposed in the paper. The principle of the method is to elevate the regression effect of heteroscedasticity model with using appropriate weights. First, a reasonable method to estimate the variance of disturbance term is explored. On this basis, the variance function of disturbance term is fitted by polynomial basis and B-spline basis respectively. Finally, weighted B-spline regression was performed on the heteroscedasticity model with the variance function as the weight. This method does not need to ignore the existence of heteroscedasticity, nor is it limited by the general assumption in the past. It has positive significance to reflect the operational performance of the new weapon system objectively which is still in the stage of scientific research and design finalization. In addition, the method can be applied in both homoscedastic and heteroscedastic populations.