Effectiveness of Attitude Estimation Processing Approaches in Tolerating Radiation Soft Errors

Abstract

This paper investigates and compares the neutroninduced soft-error tolerance effectiveness of five classical attitude estimation (AE) processing approaches that are typically embedded in inertial navigation systems of autonomous things. Results of 14-MeV and thermal neutron radiation testing campaigns indicate that all the AE approaches -implemented without protection mechanisms -can be critically perturbed by single event upsets (SEUs), recovering themselves after a few seconds if sensors' measurements are continuously provided. Moreover, Kalman filter-based AE approaches presented better effectivenesses in tolerating SEUs than AE based on gradient descent.