Adaptive Kalman filter for guided rolling projectile attitude estimation

Attitude estimation is essential for inertial navigation system. When the external acceleration is significant, great errors are introduced into measurements, which is one of the main sources of performance loss in inertial navigation system. This paper addresses two problems: one is the error calibration of inertial sensors, the other is attitude estimation during large external acceleration occurs. A maximum likelihood estimation algorithm is introduced for sensors calibration and a quaternion-based adaptive kalman filter is designed for compensating external acceleration from the residual. The adaptive algorithm is to improve the performance of classical kalman filter in various dynamic conditions. Simulation results suggest that the proposed algorithm can estimate the attitude of projectile accurately under accelerative environment, with superior performance over the classical quaternion-based kalman filter.