COMPARISON OF APPROACHES TO UNKNOWN PARAMETERS IDENTIFICATION IN GYRO DRIFT MODEL

Abstract

The paper proposes a model of floated gyro drift, used in a platform-based inertial navigation system, which takes into account temperature effects. We consider a problem of unknown parameters identification in a mathematical model of floated gyro drift. The presence of nonlinearity in the dynamics equations is expected. It is proposed to solve the formulated identification problem by Kalman-type filter application. This filter is characterized by the linearization point variation in course of solving the problem. The problem is solved for the gyro drift model which takes into account the temperature effect. The results of the model unknown parameters identification are compared using two algorithms. The first proposed algorithm is based on Kalman-like filter and the second one is based on the least squares method. The analysis of simulation results showed that the accuracy of unknown parameters identification by the algorithm based on linearized Kalman filter is commensurable with the algorithm based on the least squares method. However, linearized Kalman filter is efficient in solving the problem of identification of a gyro error model represented by combination of a few random processes.

Keywords

• error model
• nonlinear filtering
• identification
• floated gyro
• least squares method
• Kalman filter