Characterization of various IMU error sources and the effect on navigation performance

Abstract

This paper introduces a simple model for modeling inertial sensors and expands on this model to incorporate terms that model the errors associated with high dynamics. The paper uses the techniques of Allan Variance Charts, Monte Carlo Simulations, and Autocorrelation functions to characterize sensors through identification of the error statistics. The error growth of the position and heading navigation solution for two and six degree of freedom scenarios using static error parameters is developed. Analytical expressions for the error growth with wide-band noise are presented and used as an error baseline for investigating the effects of sensor drift, or walking bias. Experimental data is used to validate the error growth bounds in both scenarios. An advanced sensor model is developed for both an accelerometer and gyro with an explanation of the effects of the dynamic error parameters. A simple rocket trajectory simulation is used to illustrate the adverse effects of high dynamic sensor errors when dead-reckoning with a tactile-grade IMU. Analysis of the advanced model is concluded with an investigation into the relative effect of each error parameter on the trajectory target.