Description: It is shown that the interpretation of the unknown parameters of the signal model by random variables, as a rule, is conditional. One of the evaluation methods is considered, which reduces the problem of adaptive processing to the problem of optimal filtering. Moreover, the efficiency of the adaptive procedure constructed on this basis turns out to be essentially dependent on the degree of influence of the form and the error density parameters on the result of the estimation of the navigation parameter vector. The purpose of the article is to develop a method of adaptive processing of navigation information in conditions of uncertainty. The proposed approach to simplifying the optimal procedure of adaptive filtering, based on the discretization of the range of possible values of navigational parameters. An algorithm for estimating uncertain parameters of the navigation system error model is given. The case of dynamical change of signals models of navigation systems is considered. It is shown that the set of the model of the system error in the nominal, "normal" state with models of the system error in the state of various violations, forms a multi-structural model that can be used to construct the filter. Such a filter solves two problems: obtaining a current estimation of navigational parameters and determining the position of the navigation system on the basis of methods of the theory of statistical solutions involving the values of a posteriori probabilities of states. The method of adaptive processing of navigational information is synthesized in conditions of uncertainty of basic errors of navigation systems, in which a vector of undefined parameters is discretized through mechanisms based on recurrent target inequalities. The peculiarity of this method is that the processing of navigational information is provided when specifying undefined parameters that differ from each other not only by the numerical values of the parameters, but also by the structure. It is established that in the case of a dynamic change in signal models, it is necessary to use a pulsating filter algorithm that uses the Gaussian approximation of a posteriori density at each discrete time step, thus minimizing the increase in the number of possible sets of sequential models with increasing filtration time.
Keywords: adaptive processing, model, navigation information, navigation systems, uncertainty, filtration