Research On Comprehensive Calibration Algorithm For The Single-Axis Rotation Fiber Inertial Navigation System

The strapdown inertial navigation system (SINS) based on optic gyroscopes has the advantages of wide dynamic range, quick start. Also, it can provide autonomous navigation and positioning information. Therefore,the SINS based on optic gyroscopes become a good choice for all kind of vehicle navigation system. When certain accuracy-levels inertial sensors are adopted, the navigation performance could improve significantly by introducing the rotation mechanism while the system costs rise slightly. The Single-axis rotation inertial navigation system (SRINS) has been widely applied in various surface and underwater vehicles at present. SRINS based fiber optic gyroscopes (FOG) is becoming an important development of marine inertial navigation system. However, the constant drift of the azimuth gyro cannot be modulated by rotation, so the system navigation errors still accumulated over time. Moreover,the errors lead to the nonlinear characteristics of the error models of SRINS.A remedy to this problem is to utilize comprehensive calibration, which is a critical systematic technique to reset the navigation information and compensate the inertial sensor errors.In this paper, the single-axis strapdown inertial navigation system with optic fiber gyro is the research object. According to the problem of the nonlinear characteristic caused by long endurance systematic errors, the comprehensive calibration algorithm for the single-axis rotation fiber inertial navigation system would be researched. The nonlinear filter of the single-axis rotation inertial navigation system would be designed. The impact of the azimuth inertial measurement unit (IMU) errors on the navigation performance can be eliminated approximately and amended. The navigation parameter could be compensated such as attitude,velocity and position.The simulation results have verified the validity of this calibration approach.The principle of error compensation for SRINS is first presented. Then, based on the error characteristics of SRINS, the impacts of the constant drifting of the gyroscopes on the performance of SRINS are researched. The necessity of estimating and compensation for the azimuth gyroscope constant drifts can be found through the above analyses. According to the observability analysis theory,appropriate states of the modified UKF are selected to improve the speed and precision of the filter. Furthermore, for the state of azimuth gyroscopes and the attitude error, the observability of them is improved by adopting special maneuvering scenarios and increasing estimating time. Aiming at the problem of negative definite of the state error covariance,a modified UKF has been proposed. The filter stability is then improved by enlarging the process noise appropriately. Besides, a level damping network has been added to eliminate Schuler period oscillations and reduce the effects of gyroscope stochastic noise on the SRINS. The constant drift of the azimuth gyroscope could be approximately estimated. The modified UKF can estimate the error of the navigation information (namely velocity, position and attitude) and the constant drift of the azimuth gyroscope. Then, the navigation error and the gyro drift can be applied to the comprehensive calibration of SRINS by using feedback correction.As a result, the SRINS accuracy can be improved. Simulations results show that the position endurance levels can achieve better than that without comprehensive calibration.Also, the navigation error can decrease by 0.6n mile/48h. The effectiveness and feasibility of the calibration algorithm have been verified.