Synthesis of Civil Aircraft Control Using Empirical Data and Quantum Filtering

This study presents a new approach to solving the problem of synthesis of civil aircraft control based on methods of providing intelligent support and assessing the level of crew training. This implies the representation of flight maneuvers by canonical multidimensional non-stationary models of motion with discrete time and their matrix coefficients identified by empirical data. When performing a flight maneuver in real time, its type is recognized, and then a pattern is extracted from a specialized database to provide the correct way out of the current flight situation. This pattern is the closest to the performed maneuver in the relevance metrics, which makes it possible to calculate the sequence of control parameter vectors according to the desired change in the sequence of the aircraft states. The most interesting in mathematical terms and promising from the viewpoint of practical application, the likelihood metric for trajectories of eigenvalues has provided worse recognition results than other metrics. The study reveals the causes of such a problem and proposes a method for its elimination—quantum filtering based on the analysis of spectra of matrix transformations of flight parameters, representing the aircraft behavior. Clearing these representations from a random ‘‘noise’’ caused by errors in sampling observation, quantum filtering significantly increases the efficiency of recognition of flight fragments compared to a method relying only on the likelihood metric. Applied quantum representations are purely theoretical constructions and do not require for their implementation special computational tools, providing quantum calculations in the usual sense.