Development of software simulator of UAV flight for the studies of trajectory motion systems

Abstract

The development of unmanned aerial vehicle's (UAV) trajectory control systems is an example from a whole class of tasks that are most efficiently solved by means of computer simulation. Flight simulation allows to safely debug and study different control algorithms along with providing an opportunity to train control systems based on neural networks. There is a large number of program simulators that allow simulating a flight with a high degree of physical processes' detailization, however, the simulation speed drops significantly due to high-precision calculations. The paper considers the development of a flight simulator intended for training neural network used in flight trajectory control systems. Key features of the designed simulator are high operation speed while retaining basic features of UAV dynamics, features of distance sensors taking into obstacles in the environment.