Modeling And Compensation Of Gyro's Drifts In A Gimbaled INS

Gyroscope is the key component in an Inertial Navigation System (INS). Its drift is the biggest factor to affect INS's precision. This dissertation, based on the development task background of a project- Gimbaled Inertial Navigation System for Ship, addresses the modeling and compensation of gyro's drifts in a gimbaled INS. The main work and achievements are as following:An overview of gyro's drift is presented. Its cause is analyzed and it is categorized into three types: constant drift, slope drift and random drift. First, a method of rotating the platform during the INS's startup process is used to measure constant drift. And the effect of the constant drift on INS's precision is examined through simulation to show the necessity of proper measurement and compensation of constant drift. The slope drift in Z axis is measured using the outputs of INS after BBA calibration.Random drift is more complex than the system drift-constant drift and slope drift. Its modeling and compensation is the focus of this dissertation. Method of Box-Jenkins(B-J) which models the unstable time series after difference to make it stable, and method of Pandit-Wu(DDS) which directly model the unstable time series are both investigated. Practical modeling and comparison show DDS is of more advantages than B-J, since it not only identifies the parameters of model correctly, but also is helpful to study the components of gyro's drift. The slope term of the random drift is also identified and evaluated in this dissertation at first time.