Adaptive Alignment Method For Single-axis Rotary Inertial Navigation Systems

Rotation modulation technology is a system-level error compensation technology for strapdown inertial navigation system(SINS),which can increase the accuracy significantly during alignment and navigation.However,the system is an integral system in essence,that is to say,the errors of navigation parameters will accumulate with time.In addition,the precision alignment is a necessary stage before navigation,and its accuracy is a direct factor affecting the accuracy of navigation,while the introduction of global navigation satellite system(GNSS)as an auxiliary system can effectively improve the accuracy of alignment.In order to solve the problems of the combined system,such as Kalman filtering's poor adaptability to the environment and the loss of GNSS signals,a fuzzy multi-factor adaptive filtering algorithm is proposed.The paper focuses on the adaptive alignment algorithm of the combined system based on the single-axis rotary strapdown inertial navigation system,the major content are as follows:(1)The basic principle of single-aixs rotary SINS is studied.Besides the attitude,velocity and position error equation are derived.According to the propagation law of constant error,scale factor error and installation error of gyros and accelerometers,the modulation mechanism of the above errors by single-axis rotary SINS is derived and analyzed.(2)The method of identifying whether alignment is abnormal by setting threshold of the attitude angle error is studied to further enhance the reliability of the alignment result.Meanwhile,the basic theory,the commonly used models and their parameter estimation methods of time series prediction technology are researched.Kalman filtering can be performed according to the predicted values of the established time series model so as to ensure that the alignment process can be facilitated even when GNSS fails to work.Moreover,simulation analysis shows that the prediction through the established AR model and ARMA model can guarantee the alignment accuracy requirements during a period of GNSS failure,thereby improving the anti-interference performance of the system.(3)Aiming at the limitation of Kalman filtering,the fuzzy adaptive algorithm is introduced into the precision alignment of SINS,as well as the Sage-Husa time-varying estimator is regarded as reference information,thus the observation noise convariance matrix is adjusted in real time according to the innovation and the reference information,actually each channel independently calculates the corresponding adaptive factor during the adjustment process.Simulation results confirms the feasibility and effectiveness of the proposed fuzzy multi-factor adaptive filtering algorithm.(4)The experimental prototype of the single-axis rotary SINS has conducted multiple navigation experiments differing in environmental conditions and motion states of the carrier,including experiments on a three-axis turntable,experiments on a test vehicle,and experiments with submersibles in lakes and seas.Consequently,the experimental results shows that compared with Kalman filtering,the navigation accuracy of the fuzzy multi-factor adaptive algorithm is imporved by 4.9% to 61.8%,and even when the Sage-Husa adaptive algorithm performs poorly,the proposed algorithm can still achieve a relatively optimal alignment accuracy and navigation accuracy,which verifies the adaptability and effectiveness of the proposed algorithm for various environments.