INCREASING THE EFFICIENCY OF THE METHOD OF SEQUENTIAL DECOMPOSITION FOR SOLVING LINEAR-ESTIMATION PROBLEMS

The method of sequential decomposition can be used, along with other statistical processing methods, to minimize the amount of computation in solving estimation problems. A computationally simple method is proposed for increasing the accuracy of the algorithm of the method and for broadening the possibilities of its use. The method consists in introducing correcting terms into the algorithm that take into account the most important nondiagonal elements of the transition matrix and of the correlation matrix of the estimation errors, and also the elements of the measurement matrix.