Adaptation of discrete estimation algorithms according state model and shaping noise parameters based on the combined maximum principle

If the motion model is inconsistent with the observable state change when tracking the maneuvering target, it can lead to divergence and even failure of the estimation algorithm. Hence the development of adaptive filters is actual problem. One of the traditional variant for the filter adaptation is to use a set of identical models with different parameters. This allows taking into account the uncertainty of statistic or geometric nature for the kinematic models when describing the maneuver. However, a wide variety of the maneuver types leads to complexity for implementation of the filters built on the basis of this approach. In this paper the problem of adaptation of the discrete mathematical model to the observed system is solved as the result of structural synthesis which is obtained from the solution of inverse problem of dynamics based on the combined maximum principle.