V.M. Grinyak, A.S. Devyatisilny

Paper is devoted an actual problem of algorithms designing for support of the moving objects trajectory and steady against maneuvers, hindrances and admissions of measurements. Algorithms are bases for trajectory measurements reception in radar-tracking systems. The problem of vessels tracking by two-coordinate radar is solved in this paper. The main problem of object tracking is a stability of this process in relation to failure is. The model, reducing probability of failure and based on adaptation of traditional filtration algorithms to modeling representations of the modern theory of fuzzy logic systems is considered. Thus support is based on classical alpha- beta- filter. The approach to a choice of model from the set, based on Bayes estimations of model reliability is known. Authors essentially other approach - two-level Mamdani fuzzy logic system is offered. This system estimates vessels maneuverability and chooses a vector from a set, generated by supported algorithms. Parameters setting of fuzzy logic system can be choices by training or in the expert way (authors represent results of the last). Positive effect which gives a method offered by authors - reduction of probability of tracking failure for high-maneuver small size vessels even at their support by a strobe of small radius. Last quality gets essential weight at work VTS in the conditions of summer navigation when movement of small size high-speed and high-maneuver targets (walking boats, yachts, etc.) is saturation and intensity. Numerical modeling of the problem confirming practical usability of offered model.