Ballistic Roll Estimation using EKF Frequency Tracking and Adaptive Noise Cancellation

We present an attitude estimation method for self-guided ballistic munitions from onboard magnetometers. An adaptive feedback filter is used to remove actuator-induced magnetic disturbances whose spectrum overlaps the signal of interest. A state-augmented extended Kalman filter (EKF) is used to estimate roll rate and sensor bias from the filtered data. Experimental results demonstrate that this method is able to accurately estimate the roll without phase delay inherent in other methods.