Font Size: a A A

GNSS/INS Integrated System Model Refining And Position & Attitude Determination Using Carrier Phase

Posted on:2016-08-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y GanFull Text:PDF
GTID:1108330482479104Subject:Geodesy and Survey Engineering
Abstract/Summary:PDF Full Text Request
With high sampling rate, good continuity and comprehensive output parameters, GNSS/INS integrated system has been spread through each department of the national economy and the defense development. This dissertation mainly focuses on the theories of integrated system data processing and precise positioning, attitude determination using carrer phase based on the integrated system. The content covers GNSS/INS system data processing model refining, GNSS/RISS system filtering model establishing, INS aided GNSS precison positioning and GNSS multi-antenna attitude determination. The main works and creations are as the following:1. Current de-noising method is incapable of eliminating colored noise,so an EMD threshold de-noising method based on noise modeling is proposed. Since time-correlated colored noise is predominant, fractional Gaussian noise is utilized to model sensor errors and the model parameter is estimated by the periodogram method. Variances of the noise in Intrinsic Mode Functions(IMFs) are analyzed. Noise thresholds of IMFs are estimated through the obtained variances. To maintain the continuity of in the proximity of the zerocrossings, the interval between two zerocorssings is treated as a unified thresholding unit. Results show that EMD threshold de-noising is effective on reducing sensor errors due to its close association with proper noise model, improving navigation accuracy.2. Appropriate prior information is of great importance but usually determined according to the experience, lacking theory basis. Time-frequency methods are proposed to extract state model information from actual inertial dat, which avoids complicated tuning in utilizing Kalman filtering. For receiver clock parameter in tightly coupled integrated navigation, time-frequency methods are also used. Comprehensive comparison is made between two familiar clock state models: random walk type and Gauss-Markov type. Experimental results show that only by using Gauss-Markov type clock model together with the autocorrelation analysis of correlation time, can the variation disciplinarian of clock be represented properly. Time-Differenced Carrier Phase(TDCP) avoids the problem of integer ambiguity resolution. While it is basically a kind of relative observation in the velocity domain, suffering from error growth, thus pseudorange is still needed in another long period update cycle. The practical stochastic model of TDCP is also derived. Dual-period filtering with TDCP outperforms tradional tightly coupling with pseudorange.3. GNSS/RISS system with dual filters is designed where the heading parameter is separated with position and velocity, making both the state equation and measurement equation of the two filters are linear. Gyro information and GNSS-derived yaw are integrated through the yaw angle filter. Gyro information is used as the control input in the state model of GNSS/Gyro integration and GNSS yaw angle is made the measurement. GNSS/Gyro puts short-term accuracy of gyro and long-term stability of GNSS together. In the position and velocity filter, the state equation based on the accelerometer is derived and the sensor bias is also included. Since linearization or nonlinear filter is unnecessary, this GNSS/RISS system is effective and has good real-time performance. Attitude estimation algorithm by accelerometers is also investigated, enabling GNSS/RISS system to conduct low-cost ―3-D‖ navigation.4. Inertial Aided single frequency cycle slip detection method for real-time kinematic GNSS is proposed. Inertial aided cycle slip detection terms(DTs) are derived. Their error characteristic is analyzed comprehensively. The error DTs is influenced by the error of INS positioning increment, and for each satellite it is related to the angle between the line of sight vector of this satellite and that of the base satellite. It is proposed that two group of DTs can be formed by selecting two different base satellites and the used together. The detection threshold is estimated in a moving window, where the small-value DTs are eliminated to reflect the effects of INS error. The threshold has strong self adaptability.5. Usually it is considered that INS whose position accuracy higher than 0.5 m can improve ambiguity resolution. The error propagation property of INS is derived based on INS error equation and then stochastical model of INS virtual observation is decided. This model is related to the losslock duration, reflecting the actual accuracy of the virtual observation and balancing INS and GNSS information. Ambiguity resolution can still be improved when INS error reaches 4 m. Even when INS error reaches 10 m, GNSS itself will not be injured by INS.6. A single-epoch ambiguity resolution algorithm without the process of searching is proposed. Four basic satellites are selected, whose integer ambiguities are estimated by utilization of INS position.(1,-1) and(-3,4) ambiguity combination is formed and the ionospheric error of(-3,4) combination is polynomial fitted. After the systematic influence such as the ionospheric error is compensated, the ambiguity combination can be fixed by simple rounding. The ambiguities of each frequency can be determined by the combination and returned to resolve the receiver position. Based on this position, the ambiguities of all other satellites can be directly calculated and fixed by rounding. The algorithm is highly effective since it avoids ambiguity searching.7. In baseline-attitude calculation approach, virtual observation constructed by baseline information is used. Since the accuracy of initial baseline value is poor, the first order Taylor series expansion equation is derived to decrease linearization error. Under the condition of multi-baselines, the constraints between baselines are used to connect the ambiguities from different baselines, greatly improving success rate of ambiguity fixing.8. Attitude determination directly from carrier phase observation based on adaptively robust Kalman filtering is proposed. The state models for attitude estimation with and without external angular rate sensors are both erected, combined with the observation model based on misalign ment angle or multiplicative quaternion error. The attitude parameter is estimated by adaptively robust filtering, in which the adaptive factors of ambiguity and attitude error are decided respectively following the idea of multi adaptive factor filtering. The factor of attitude is determined by a three-section function containing Ratio. Adaptively robust filtering makes full use of constraint and historical information, fusing them in the calculation of float solution. As the accuracy of float solution and the structure of covariance matrix are improved greatly, the fix solution can be searched efficiently using LAMBDA method merely, perfectly fulfilling the real-time requirement. By seclecting a very small factor for ambiguity, cycle slip detection process can be removed in the frame of filtering, enhancing the filexibily.
Keywords/Search Tags:EMD(Empirical Mode Decomposition), RISS(Reduced Inertial Sensor System), Single-Antenna GNSS/Gyroscope Integration, Inertial Aiding, Cycle Slip Detection, Integer Ambiguity, Multi-Antenna Attitude Determination, Adaptively Robust Filtering
PDF Full Text Request
Related items