System safety of human centrifuge and solving angular velocity of main arm with artificial neural network

In high-performance military aircraft, high centrifugal forces result in sharp turns and such emergency maneuvers significantly affect the pilot's physiological condition and performance. The human centrifuge system is a dynamic flight simulator used to generate artificial G forces realistically. One of the objectives of this study is to underline the significance of system safety of human centrifuge systems. System safety should be considered well during the system design to prevent possible problems during training. Failure mode and effects analysis is suggested for making the system reliable. In this study, new solutions suggested for obtaining better results from previous research for calculation of the angular velocity of the arm used to move the system to provide G forces required in pilot training. The differential equation used to get angular velocity is described to show where the problem arises. Various well-known methods including Euler, classic Runge-Kutta, Cash-Karp, Dormand-Prince, Fehlberg, Artificial Neural Network, Curve Fitting, Radial Basis Function Network are used for solving angular velocity of arm. A mathematical model created by using neural network provides a solution having good generalization for G commands generated using smooth step algorithm. Runge-Kutta family is a good method to solve angular velocity for increasing G commands. However, Runge-Kutta methods don't give satisfactory results for decreasing G commands, therefore, this study mostly focused on solving decreasing G commands.