Adaptive Minimum Mean Square Error Forecast of Missile Trajectory Using Stochastic Difference Equation Models.

摘要

Data projection (forecasting) methods are illustrated using missile data supplied by Quality Evaluation Division of White Sands Missile Range. The process of model building making iterative use of identification (using the autocorrelation function) fitting (maximum likelihood estimation) and diagnostic checking (analysis residuals) is illustrated in the building of an appropriate stochastic difference equation model. It is shown in detail how all the following may be calculated directly from the model: the projection (forecast) function, the memory function, the error of the projected value at any lead time; and the updating of the projection function.