Experiment on signal filter combinations for the analysis of information from inertial measurement units in AOCS

Abstract

The FDIR software subsystem may be part of the attitude and orbit control subsystem, AOCS. The AOCS quite often includes inertial navigation sensors being physically implemented by accelerometers and gyroscopes which provide electrical signals to the AOCS on-board software (OBSW) which, in its turn, generates the commands to the control actuators. In general, hardware like sensors and actuators present nonlinearities which sometimes make it difficult to properly interpret the output signals. In the scenario of space applications, filters are used to eliminate noise and to increase the reliability for the correct interpretation of those signals. In this paper we present a collection of filters used in inertial navigation subsystems enabling the fusion of data from sensors. Fundamentally, the filters are composed of the Kalman filter in its derivations. The filters can be used for state estimation of a system as well as for noise filtering. In this work the filters are configured with respect to their different orders of execution, their sampling rate, and their cutting-off frequency. The filter configurations can be changed by software so as to allow a flexible structure that can be adjusted for the best quality of output signal and consequently the best analysis of the satellite behaviour. The main purpose of this paper is to test the algorithm that combines several signal filters considered in this study. To accomplish this goal we developed an experiment encompassing an accelerometer and a wireless communication system so as to provide input signals to be filtered by the filtering algorithm.