A.A. Konovalov

An issues of the single target track formation algorithm design for the Multi-Site Radar Complex are discussed: a choice of tentative track formation criterion, tentative gate shaping, calculation of the resulting target position estimation based on the measurements from several radars. In MSRC we have to use track formation criteria like «n measurements at time with probability p», or Expression for the probabilities of equal n measurements gaining are derived based on the statistical properties of the random point process of the measurements from the same target but several radars. Further we obtain an expression for the probabilities of not less than n measurements gaining at the time which is used for tentative track formation criterion. Conventional tentative track gate is ring-shaped with radiuses proportional to the minimal and maximal expectant target velocities. But in this case gating probability depends on the target-s range and velocity, radar measurements errors and may take on a small values about 0.5. So gating technique have to take in account a radar measurements errors. We propose two-stage gating technique with conventional ring-shaped gate on the first stage and elliptical gate (the same as correlation gate in the tracking data association methods) on the second stage, needed when measurement not falling into the ring gate. Elliptical gate builds up around second measurements; it-s size depends on the correlation matrix of both measurements. This technique provides lower false track initiation probability than more simple method based on the external ring extension. After gating we have a set of true and false measurements. More than one of true measurements (from different radars and several scans) may be in this set. True measurement selection is needed. Proposed selection method is based on the velocity vector calculation for the each pair «first measurement - strobed measurement», and pairs with nearly the same velocity vector considered as a true measurements. Fused target position estimate calculates by means of appropriate track filtering algorithms.