This paper presents a calibration method based on a multiple actuator function (MAF) to improve the navigation performance of the inertial navigation system (INS). The navigation performance of the INS can be improved by utilizing a compensation function. Existing calibration methods model the compensation function based on calibration coefficients obtained by indirect calibration. In indirect calibration, the calibration coefficients are calculated using acceleration errors. However, errors such as random walks, white noise, and bias instability can affect the precision of the calculated calibration coefficients. These errors can degrade the accuracy of the calibration coefficients and the compensation function. To overcome these limitations, the proposed method models a compensation function based on the MAF. The accuracy of the compensation function is improved by the accurate actuator angle and actuator position of the MAF. Unlike indirect calibration, the precision of the MAF is improved exclusively by navigation performance. The accurate actuator angle is calculated by adopting gradient descent and Q-learning, and the accurate actuator position is calculated by adopting the Bhattacharyya coefficient. The accuracy and precision of the proposed calibration method is evaluated by static-state tests and vehicle tests. The results show that the proposed calibration method is a valid approach to improve the navigation performance of the INS.
展开▼
