Collective Object Recognition when Prioritizing Monitoring Tools

In accordance with the principles of structural and system monitoring of the situation, the process of processing heterogeneous data includes a procedure for identifying targets, the de-masking features of which appear in various information signals. Identification consists in making a decision on belonging of a group of de-masking features to one object on the basis of comparison of its angular coordinates. When using the ideal observer criterion, objects are identified for which the estimates of angular coordinates are grouped with the largest values of the conditional probabilities of the identification situations. A priori uncertainty of the same type of features, due to errors in detection and measurement of their values in the monitoring equipment, is prevented by finding estimates of maximum likelihood with dispersions, established operational characteristics of the receiving and measuring equipment. Algorithms for filtering parameters of trajectories of accompanied maneuvering objects have been developed. To minimize medium-quadratic errors in determining identification parameters, the article builds algorithms for extrapolating the movement routes of objects with the task of correcting noise acceleration and replacing the results of filtering coordinate estimates with measured values, when determining the fact of changing the real trajectory relative to the predicted. Due to the integration of information from various sensors, the list of features of objects is expanded, allowing for a more detailed classification of their types and analysis of current states.