Individually-adapted Control and Support of Pilot's Control Actions Based on Neural Network Models

The report presents the concept of intelligent monitoring and support for pilot control activities when performing an aircraft landing modes. The approach uses a neural network model of the system “aircraft-pilot‿. It is supposed that the model is unique for every pilot. This model is continuously updated in the process of pilot activities according to the results of all previous flights, and reflects the individual manner of controlling the aircraft specific to this very pilot. This approach provides the greater forecast accuracy of the aircraft path while performing a landing. The forecast is used to generate the support signals to the pilot in order to increase the landings accuracy. The report also presents the results of testing the developed neural network model at the set of data obtained using a modern aircraft simulation facility. The report also discusses the use of the individual neural network model of the pilot for generating optional indicating signals to diminish course and glide path deviations while performing the landing.