Engineering Physics Annotation << Back IDENTIFICATION AND ADJUSTABLE INERTIAL NAVIGATION A.V. CHERNODAROV, S.A. IVANOV This paper is devoted to the problem of increasing the reliability of error estimation of inertial measurement units (IMU) with the help of an extended Kalman fi lter (EKF). This problem is related to the inadequacy of noise models of inertial sensors and real processes. Gyroscopes and accelerometers are considered as inertial sensors. In known works inadequacy of the models is compensated by the corresponding tuning of the parameters of the EKF by the innovative sequence. However, in real operating conditions, due to errors of external measuring means or lack thereof, such tuning is not always possible. There are also approaches to the estimation of the statistical characteristics of inertial sensors using the Allan method. This method allows us to estimate the stability of errors on the moving intervals of averaging. However, such approaches are not related to tuning up the EKF parameters. Therefore, their use does not ensure the adaptation of the EKF in real operating conditions. The scientific originality of the proposed work is connected with the addition of procedures for tuning noise models of inertial sensors to known EKF adaptation algorithms. It is proposed to perform the adjustment of models on the basis of structural-parametric identifi cation by means of correlation processing of error estimates of sensors. Such processing can be performed both in real time, and according to the data of on-board recording devices. The developed algorithms allow to take into account the change in the accuracy and dynamic characteristics of inertial sensors through the corresponding coeffi cients of the IMU errors. To implement the proposed identifi cation technology, the errors of inertial sensors should be included in the vector of the estimated parameters. The paper presents the results of experiments confi rming the expediency of identifying models of noises of inertial sensors during operation. Key words: inertial navigation system, sensors, noise model, identifi cation, Kalman filter. DOI: 10.25791/infizik.06.2018.046 Contacts: E-mail: chernod@mail.ru, E-mail: srpremier@mail.ru Pp. 29-37.