APPLICATION OF GENETIC ALGORITHM FOR MODELING AND IDENTIFICATION OF UNMANNED AERIAL VEHICLE MOTION

This paper presents a comprehensive study of system identification of an unmanned aerial vehicle (UAV) of an airplane type based on experimental data collected during longitudinal flight. The study focuses on identifying and deriving mathematical models of the UAV in the pitch control channel without feedback or controllers. Using MATLAB System Identification Toolbox, linear mathematical models are developed as integral-differentiating relations of the second and third order. Another approach obtains a linear model using a genetic algorithm by optimizing the structural scheme of the UAV’s longitudinal short-period motion. The paper provides a comparative analysis of these models, assessing their accuracy, computational complexity, and applicability in control design on stationary PCs and onboard computers in adaptive systems. The results highlight trade-offs between modelling approaches, offering valuable insights for researchers and engineers in UAV system identification and control.