Research On Estimation And Compensation Of Error Model Parameters In High-precise SINS

The thermal drift errors of inertial device and the sequential start-up errors of INS are the significant errors in the high-precise inertial navigation. In order to satisfy the accuracy in the long-range marine navigation and airborne gravity survey applications, the following errors can be compensated to measure the more accurate angular incremental quantity and velocity incremental quantity. First, the temperature drift errors of RLG bias are needed to be compensated. Second, the temperature drift errors of scale factors, misalignments and biases of quartz flexibility accelerometers are needed to be compensated. At last, the on-line start-up errors of gyroscopes and accelerometers are needed to be compensated.The main achievements of this thesis can be summarized as follows.1. A multi-position estimation method for the RLG biases based on the inertial reference-frame computation is proposed, which is restrained by the Earth rotation velocity. Thus the accurate specific force measured by the triaxial accelerometers in the conventional multi-position north-seeking procedure is not needed any more. Then the multi-position calibration of RLG biases can be successfully implemented within the partial observation space supported by the inaccurate double-axis turntable. Then the above two kinds of estimation accuracy of RLG biases is equivalent by the triaxial turntable calibration test in the room-temperature condition. Then the discrete thermal calibration of RLG biases is effectively implemented in the multi-temperature environment supported by the two-axis incubator. Finally, the three-order regression model of RLG biases is successfully established to obtain the high-precise angular motion measurement in any thermal condition.2. Based on the damped least squares, a multi-position estimation method for the linear model parameters of the triaxial quartz flexible accelerometers is proposed, which can avoid the turntable error effects on the estimation accuracy. The proposed method has the equivalent estimation accuracy as the gravity-equation-solving algorithm illustrated by the simulation and test results. Then the relative attitude parameters between the gyroscope triad and the accelerometer triad are successfully estimated by the constrained quaternion optimization method. The discrete thermal calibration of accelerometers is effectively implemented in the multi-temperature environment supported by the two-axis incubator. Finally, the three-order regression model of accelerometers is successfully established to obtain the high-precise linear motion measurement in any thermal condition.3. A rapid thermal calibration method of accelerometer nonlinear model parameters in the start-up procedure is proposed, which utilizes the sequent attitude reference observation measured by the triaxial gyroscopes. Thus the calibration cost and consumption is effectively reduced. Due to the deficiency of more model errors brought by the least-square identification to the thermal model parameters of accelerometers, an DWO-based optimization estimation method is proposed to identify the corresponding thermal parameters. The more accurate gravity measurement by the proposed method is effectively validated by the start-up procedure.4. A 15-state Kalman filter model is proposed with the observed quantity of position errors. Then the systematic on-line calibration of the high-precise SINS is effectively implemented by the vehicle maneuver. The biases of horizontal gyroscopes and accelerometers are then successfully estimated during the time span of three hours. Finally, the start-up errors of inertial device are apparently compensated to improve the navigation accuracy of high-precise SINS during the short period.