Method of Combining the Synthesis of Program Control with Homing Methods for the Problem of UAV’s Control in Conditions of Multifactorial Uncertainty

Abstract

The method of statistical synthesis of algorithms for constructing training samples of neural network support of UAV’s control process is developed. The statistical evaluation of UAV’s effectiveness by means of probability of achievement of the purpose on condition of the maximum and minimum of criterion of regularity is presented. The algorithm for generating control actions according to the target environment is implemented in the form ofa trained onboard neural network.