Strapdown Algorithm And Filter Technique For SINS/GPS Integrated Measurement

For the requirement of precise navigation practice in complex environments, two study aims are determined to:(1) investigating how to design the efficient and precise SINS navigation algorithm in complex environments, and mainly studying how to compensate the non-commutativity error of SINS navigation computation in complex environments such as a severe vibration or maneuver environment; (2) studying the nonlinear data fusion for SINS/GPS integrated system, and seeking a precise and robust non-linear filter algorithm and data fusion policy applicable to SINS/GPS integrated system in complex environments, in order to supply the support in theory and technology for the precise navigation and location in complex environments.Through analyzing the traditional strapdown attitude computation structure and algorithm design method, it is found out that the tradition coning correction algorithm structure has obvious defects, that the traditional coning algorithm based on the compressed or uncompressed correction form has problems of low accuracy or efficiency in complex environments. Thus, optimizing the coning correction structure is as the breakthrough port, to realize the optimal design of precise and efficient strapdown attitude coning algorithms in complex environments.First, a constraint coning correction form is presented in this paper. This coning correction form can be considered as the simplified version of the uncompressed coning correction form with some constraint relation. Based on various optimization design principles, (In a sense, coning correction residual is minimized in the classical coning, an expected discrete coning, a vibration, or a maneuver motion.) the coefficients of the constraint coning correction form are optimized, then the constraint coning correction algorithm is achieved. Compared with the uncompressed coning algorithm design, the constraint coning algorithm design needs an additional consideration of some constraint relation. In theory, the constraint coning algorithm is more accurate than the uncompressed, when both are with the same algorithm throughput.Meanwhile, a half-compressed coning correction form is presented also. This coning correction form can be considered as the simplified version of the uncompressed coning correction form with some severer constraint relation. The coefficients of the half-compressed and the traditional compressed coning correction forms are with the same free degree. Between both is some simple relation, which makes that the half-compressed coning correction algorithm coefficients can be directly derived from the known compressed coning correction algorithm coefficients. Thus, the design process of half-compressed coning algorithm is simplified. Compared with the compressed and uncompressed coning algorithms, the half-compressed coning algorithm is temperate in maneuver accuracy and algorithm throughput.For the angular rate-based SINS, a novel angular rate-based coning correction form is presented in this paper. This coning correction form includes angular rate sample cross-product, angular increment sample cross-product, and the cross-product of angular rate and increment samples. Based on the novel correction form, a class of angular rate-based coning algorithms are designed in the classical coning, discrete coning and vibration motions. Compared with the traditional angular rate-based compressed coning algorithm, the new algorithm is more accurate in whether coning environment or maneuver environment.For the strapdown inertial navigation velocity computation, the constraint sculling algorithm and the half-compressed sculling algorithm are directly designed by converting respectively the known constraint coning algorithm and half-compressed coning algorithm, on the basis of the dual principle between coning algorithm and sculling algorithm, and some equivalent conversion method. For the strapdown inertial navigation position computation, a constrain rotating and scrolling correction form is presented. Based on this correction form, a constraint rotating and scrolling correction algorithm is designed using the frequency Taylor method. Compared with the traditional compressed rotating and scrolling correction algorithm, the new algorithm is superior in maneuver performance and algorithm throughput.For the application demand of SINS/GPS integrated navigation data fusion, integrated system error analysis, SINS initial alignment, SINS calibration, and so on, a high-order SINS multi-error coupling and propagation model is established, by analyzing the way of various error sources into SINS/GPS integrated navigation system. In addition, the state equation and measurement equation of SINS/GPS tightly integrated navigation system are also built.By analyzing the characteristics of EKF, STKF and SHKF algorithms and IMM data fusion policy, an IMM data fusion policy-based filtering algorithm for SINS/GPS tightly integrated navigation is presented. This filter includes three subfilters:one STKF and two suboptimal SHKF s, and each is based on an independent model. Compared with the EKF, STKF and SHKF algorithms, the IMM-AKF algorithm is more robust and of good filtering performance under uncertainty condition.