The article pays special attention to increasing the requirements for the security of objects of special importance, the state border, water and air transport in the face of growing high-tech terrorist threats. The achievement of these requirements is inextricably linked with the provision of quality indicators of technical means of remote monitoring of dangerous objects, the complexity and intellectualization of location systems, sensor networks with the functions of network-centric interaction. The use of electronic countermeasures by the enemy, the rapid change of scenarios of combined interference determine trends that are aimed at increasing the probability of obtaining information in location systems, adaptation to the interference of various origins at short intervals. Lack of reliable information about the presence of the object can lead to unacceptable environmental and material losses, and false alarms – to unjustified additional costs. In this regard, an important area is to provide the most reliable information in the face of interference of various origins. The urgency of the problem is related to the search for a compromise between costs (cost and time of development) to build adaptive detection and recognition systems, and on the other hand, efficiency (simulation probability, implementation complexity), which is achieved by developing optimization criteria. At the same time, special attention is paid to ensuring high-quality indicators of detection and recognition in the conditions of changing and a priori unknown noise situation.
An approach and criteria for the synthesis and analysis of digital adaptive detection and recognition systems have been developed. The criteria are based on the decomposition of the average risk and the probability of errors in the decisions made, due to: the final size of the system, the size of the training sample and the bit count. This approach allows you to consistently solve the problem of synthesis and analysis of optimal systems.