Influence of the sensor error model on stability and controllability of a simple autopilot of a small unmanned aircraft

Abstract

The paper deals with the creation of a three-axis error model of the angular velocity sensor using the analysis of the ADXRS150ABG sensor installed in the IMU MNAV100CA. The created error model takes into account the basic errors affecting the sensor, which are Angular random walk, Bias drift, Misalignment error and Non-orthogonality error. Analysis of these noise mechanisms was performed using Allan variation. The authors also discuss the creation of a model of longitudinal movement of the aircraft, where they verify the correct functioning of the error model of the sensor. The resulting simulation proves that the error model destabilizes the longitudinal control of the airplane and so it is necessary to design filters. These filters remove the noise components from sensors information. By removing these components, it is possible to increase the stability and controllability of the airplane.