 ### Modeling Stochastic Errors of Inertial Sensors

• If you have difficulties in understanding the specification sheet values of inertial sensors or obtaining error model parameters from the Allan variance curves, this introductory note on stochastic errors of inertial sensors may help you.
• Check this link for an informal critique of Allan variance method.
 allan2markov_v000.m Given the (peak) value of the Allan standard deviation, computes the associated 1st order markov parameters. This may be used to approximate the flicker noise as AR(1). arma2allan_v000.m Computes the theoretical allan variance for arma (p,n) (p < n)(including rw) models. You may use this functin to verify how close your approximate markovian error models is to the actual (observed) Allan variance. arma2cov_v000.m Given an arma model, computes the exponential coefficients of theoretical correlation of the model. (You must note that the correlation of arma models are always in the form of exponential functions) cp_allan_var_v000.m Computes the allan variance of the inertial sensor outputs. Most probably this is the only function that you may need in this section. You can extract the stochastic error model coefficients from the AV figure computed by this script. cp_wn_pow_v000.m computes the disturbance power of the random walk and white noise processes. Example_conf.m An example script which demonstrates the use of modeling functions in this section. exp2markov_v000.m A helper function for the example script exp_approx_v000.m Given some exponential functions, computes a single approximate exponential. (Helper function for the example) exp_approx1_v000.m Similar to exp_approx exp_approx2_v000.m Similar to exp_approx exp_approx3_v000.m Similar to exp_approx expo_fit_v000.m Given the autocorrelation data, fits an exponential function lvl_filter_v000.m DCT based Haar filter implementation. (See the example for the use of this function in correlation computations)