A Software in the Loop (SIL) Kalman and Complementary Filter Implementation on X-Plane for UAVs

Abstract

The paper presents a software in the loop (SIL) sensor study in simulation environments for traditional Kalman, linear and nonlinear complementary filters, which are derived, tested and implemented on a fixed wing UAV for attitude estimation (pitch, roll and heading angle). An overview of the SIL setup environment between MATLAB/Simulink and the X-Plane flight simulator is given. Kalman filter design in Simulink utilizes a state-space model of the UAV dynamics, while complementary filter combines accelerometer output for low frequency attitude estimation with integrated gyro output for high frequency estimation. Simulation results are provided and discussed under both Gaussian and uniform noise, highlighting the convergence of the designed estimators. It is also shown that the estimator following the nonlinear complementary framework yields a better match to the dynamic evolution of the actual attitude angles of the vehicle over time.