Study On The On-line Calibration On The Key Parameters Of The Inertial Navigation System And The Error Compensation Technology

The inertial navigation system(INS) has been widely used in the fields of sea, land, air and space because of its independence, security and accuracy. However, the accumulation of the errors of inertial instruments causes that the precision of the system decline, the inhibition of the errors is a primary issue of navigation study. On the premise that the INS is not removed, it enhances the navigation accuracy of INS by the following methods. Selecting a proper maneuver to make sure that all the errors are observable, obtaining the observability degrees to judge whether the evaluation results are reliable or not, using Kalman filter technology to measure the IMU errors, compensation on IMU via the reliable measurements. The accuracy of the calibration results briefly depends on the followings: 1.Designing the suitable calibration maneuvers; 2. Acquiring all the observability degrees of INS errors; 3.Using Kalman filter to obtain the optimal estimations. This paper carried out the following research based on the above questions:Firstly, considering the influence of INS maneuvers on observability, which reveals the convergence of estimations. To ensure all the errors are observable, inverse-observability was put forward for designing maneuver. Furthermore, the criterion for designing INS calibration maneuver has been proved.Secondly, to judge the reliability of the estimations, after summarizing the existing methods a novel approach based on solving constraint optimization problems for observability degree has been proposed, the efficiency and the accuracy of the new method was already proved, and a simulation of new method was also carried out to verify the practical efficiency of new method. Thirdly, Kalman filter is used to accomplish the measurement of IMU errors. However, to avoid the side effects of the complexity and uncertainty of practical environment, an adaptive Kalman filter which is based on the combination of system output has been proposed. It made full use of system outputs to evaluate the unknown parameters of system model. This method was implemented on INS and the accuracy of the filtering results have been improved. Lastly, the efficiency of the whole scheme was also proved via the practical experiment. Compared with the traditional methods, the navigation accuracy has been improved significantly.