Approach to conditioning improvement in scalar calibration problem for three-axis accelerometer module using visualization of measurement efficiency function

The paper proposes a way to optimize a set of calibration angular orientations of accelerometer module using efficiency function visualization for stationary based calibration to increase an accuracy of estimated error model parameters. It includes an analysis of mathematical model of three-axis accelerometer module measurement errors. The method improves an estimation accuracy of analyzed error model parameters. The error model includes following factors: angular errors in sensor alignment in module frame of reference, deviations of sensor scalar coefficients and sensor biases. Measurement efficiency function characterize an impact of each newly made measurement on overall problem conditioning and depends on module angular position relative to calibration reference vector. By determining minima points of shown function, it is possible to form an optimal set of angular positions for calibrated module, which allows to achieve better conditioning of calibration problem. These minimal points are determined via optimization algorithm. Due to function complexity the visualization is necessary to find and set initial points for searching. The approach is verified by computer simulation which shows that optimal set of angular module positions (optimized set), formed by presented method, improves estimation accuracy of considered parameters in error model in presence of errors in angular positioning of module during calibration process, in comparison with non-optimized set. © 2020 National Research Nuclear University. All rights reserved.