RVM-based prediction and compensation method for the long-term stability of INS system parameters

Abstract:

To improve the long-term stability performance, reduce the manual calibration costs and enhance the use efficiency of inertial navigation systems, we propose a prediction and compensation method for the long-term stability of inertial navigation system parameter based on correlation vector machine, in which we choose the mean and standard deviation as the performance indicators. For the mean parameter with a significant change with time in the law, we establish regression modeling for longer storage stability parameter by RVM method, and carry out the performance prediction and calibration parameters compensation for the parameters stability of the shorter storage. The paper presents the modeling forecasting and parameter compensation for the long-term stability of the accelerometer scale factor of the important parameters in the inertial navigation system, the parameter stability mean performance for an interval time of 6 months has improved by 50.90%. This result implies that this method can replace the manual calibration and the use efficiency of inertial navigation systems, and verifies the effectiveness of the proposed method.