A.A. Ivankov

Improving the performance of the aircraft brought to a more complex level of the problem of simultaneous tracking of airborne targets. Used in modern airborne radars (radar) tracking algorithm does not fully meet the requirements for precision coordinate measuring intense maneuvering targets and the sustainability of their support. The measurements coming from a group of tracked targets, as a rule, you must identify. In modern air combat, which is intense maneuvering, often the entanglement of the trajectories, which greatly complicates maintenance. To avoid this, it is necessary to develop a method to identify the trajectories of several air targets using information from a single radar. Location of targets relative to radar, especially after the intersection of their trajectories is largely random, therefore it is expedient to use the principle of weighing the estimates of phase coordinates, based on the Bayesian approach. Providing high accuracy coordinate measuring goals can be achieved by adjusting the coefficients of the model state under the real parameters of target movement, or by adapting to the specific conditions of operation. The most effective way of adaptation is the adaptation based on a parametric model identification condition. The above data are the basis for developing a method of adaptive probabilistic binding trajectories (AVPT) movement of air targets. It will allow for the synthesis of algorithms of binding measured coordinates with the full uncertainty of their compliance with any of the ob-served targets, if accompanied by a few goals a radar, used to describe the dynamics of aircraft nonlinear adaptive model. Essence AVPT is the following. To bind with a certain probability measures the coordinates of the trajectory i-th goal, where i = 1,2 ... m, we introduce a random vector parameter q taking m possible values: q1 = [10 ... 0], q2 = [01 ... 0] ..., qm = [00 ... 1]. Proper identification is reduced to finding a posteriori probabilities P(qS/z), where s = m!  The number of possible states of the matrix qт=[q1, q2, ? qm]. These probabilities are used as weights in calculating the estimates of phase coordinates of tracked targets, and when binding measurements. Simultaneously, the iterative adaptation of the model state to the conditions of operation of the tracking system (movement of air targets) by identifying the parameters of the model. For this filtering algorithm is synthesized based on the method of invariant immersion, using a nonlinear adaptive model. The result shall be identified as a state model and observation model. At the same time monitored the trajectory of motion and prevents entanglement goals in passing their paths simultaneously through a single resolution element tracking radar system, which plays an important role in the conduct of maneuver battlefield.