Vu Chi Tkhan’, D.A. Milyakov, B.G. Tatarsky
Vu Chi Tkhan’, D. A. Milyakov, B. G. Tatarsky
Kalman filtration algorithms considering actual parameters uncertainties in dynamic model of observation process and statistical parameters of observation measurement errors are considered. The mathematical apparatus of artificial intelligence with fuzzy sets for problem solving was proposed. The solving is choosing the amplification coefficient for Kalman filter on the basis of the dynamic analysis. The dynamic comprises the disparity variances for various conditions of observation and dynamic model parameters for observation process. The compositional inference rule for choosing the adequate mean of amplification coefficient is offered. This rule consider a priori information about possible means of models parameters and conditions of observation and also actual observations. A priory intelligence are presented in qualitative form and compose the intelligent database. The database is presented as complex of condition-action rules. The transition from fuzzy data obtained during fuzzy inference algorithm operation to explicit form is realized using the center of balance method. The presented approached attitude reduce time of the amplification coefficient for Kalman filter solving.