Methods and Procedure for the Nonlinear Dynamic Characterization of Inertial Sensors Under Diverse Operational Conditions

Abstract

Inertial Measurement Units or Navigation Systems (IMU, INS) generally comprise sensors, namely accelerometers and gyroscopes, in many configurations for aerospace Guidance Navigation and Control applications. The sensors are generally modelled and tested for only a few parameters, limiting the behaviour knowledge under adverse environmental conditions. Using established and standard testing, such as vibration shaker and rotation table facilities, we developed a methodology to model the nonlinear parameters of sensors, especially those under vibration, high acceleration, and varying thermal conditions. The method records the data in the time domain, establishes the ground truth through sensor fusion, corrects misalignments and identifies ARMAX and NARMAX models using recursive Unified Least Squares (ULS) method. These identification methods are adapted and use a proposed model selection strategy based on the Error Reduction Ratio (ERR), the t-statistic and the Akaike Information Criterion (AIC). The simulation results validate the method’s concept and provide insight into the reach of the models derived for accelerometer sensors.