Reducing type I errors in aircraft contour recognition using collective intelligence of unmanned aerial vehicles

This study work presents the problem of reducing type I errors in aircraft contour recognition and offers a solution for it. The study models a distributed intelligence of a group of unmanned aerial vehicles simulated in their onboard computers using pretrained neural networks. It provides a definition of an evolutionary solution matching method with a theoretical description based on genetic algorithms, Condorcet's jury theorem, and the Rasch model. The study demonstrates conditions significantly reducing the probability of wrong decisions. It offers and tests a two-level hierarchy of collective intelligence based on collective application of evolutionary matching using neural networks as intelligent agents.