| As important as knowing position,velocity and time information of the vehicle,attitude determination is one of the key technologies in aviation,aerospace,navigation and land,and is a hotspot in the navigation and control field.In the recent years,Global Navigation Satellite System(GNSS)has gained more and more attention in the attitude measurement because of its high-precision,all-weather and long-term stability.Attitude can be estimated by multi-antenna GNSS array or integrating single/dual antenna GNSS and other sensors,for example,GNSS and inertial measurement unit(IMU)are fused for attitude estimation.This article will focus on multi-antenna GNSS and singleantennas GNSS/IMU integration attitude estimation algorithm.The specific contents are as follows:(1)Taking into account the complexity of the vehiclar kinematic environment,ambiguity resolution strategy is as follows: considering the Bootstrapping success rate,the ambiguities after decorrelation is divided into reliably fixed part and unfixed part.If the former passes Ratio test and can be fixed as integers,at the next epoch,it will be regarded as constants,while the corrected ones of the latter,along with the corresponding covariance matrix,will be treated as prior information or pseudomeasurements to participate in the calculation.Otherwise,both parts will be processed in the latter way.That is,the historical information of ambiguities is fully utilized and the partial ambiguity resolution method is used,which effectively improves float ambiguities' accuracy and the fixing success rate.In addition,multi-antenna GNSS attitude determination model based on baseline coordinate vector and the attitude angle is derived in detail,and the complementarity of the two algorithms in terms of parameter redundancy and model strength is summarized.Considering the abovementioned ambiguity resolution strategy,a switching-parameterization GNSS attitude determination algorithm taking the number of fixed ambiguities as index is proposed.Compared with the traditional method,its accuracy in terms of pitch and roll angles is improved by 28.06% and 14.45%,respectively.(2)Use Euler angles to represent attitude information and establish its kinematics model,and then introduce it into multi-antenna GNSS attitude estimation.While considering the time-varying and uncertainty of attitude motion,adaptive Kalman filtering algorithm based on the maximum a posteriori estimation principle is used to adaptively adjust system noise covariance matrix of the partial state vector.In addition,in order to accelerate the filtering convergence and improve the estimation accuracy,the filtering starts after the initial value is accurate enough.For ambiguity resolution,the least-squares ambiguity decorrelation adjustment method is used,and hence the optimal value of Euler angle is output.Experimental results show that the proposed method has better accuracy,stability and adaptability.Compared with the traditional method,its accuracy is improved by 19.43%,16.62% and 27.84% in yaw,pitch and roll angle,respectively.(3)Aiming at the problem that the accuracy of IMU attitude determination is affected by acceleration,a filter algorithm that uses instantaneous acceleration information calculated by the Time-Differenced Carrier Phase(TDCP)measurement to assist IMU attitude estimation is proposed.The above algorithm processes the displacement increment represented by TDCP in a polynomial modeling manner.A computationally efficient rank one update algorithm is used to obtain instantaneous acceleration information,which is deeply integrated with the angular velocity and specific force information of IMU.At the same time,in the filtering algorithm,the details such as attitude representation,attitude kinematics modeling,IMU error modeling,state estimate resetting and the covariance resetting are considered.Experimental results show that the proposed algorithm significantly improves the accuracy of IMU attitude estimation. |
PDF Full Text Request