Black-box Identification of a Robotic Flight Simulator

Optimization of motion cueing algorithms is a relevant topic in flight simulation industry. To achieve this goal, a representative model of the motion platform is needed. With this intent, this work presents a black-box approach to identify a model for the SIVOR flight simulator. The model receives position/orientation inputs in Cartesian space and streams out accelerations/angular speeds measured on the pilots' head. A comprehensive experimental analysis is carried out using a dataset based on isolated positional and rotational inputs. A frequency domain analysis is performed to evaluate signal measurement noise and to determine the model structure. As a result, a combination of continuous time linear transfer functions were identified to represent each of the direct and cross relations mapped throughout the process.