GNSS-A水下定位的动态非线性Gauss-Helmert模型及其抗差总体卡尔曼滤波算法

Dynamic nolinear Gauss-Helmert model and its robust total Kalman filter algorithm for GNSS-acoustic underwater positioning

GNSS-声学组合式观测是确定海底控制点位置的重要手段,但会受到声速不确定性、海面平台定位偏差等误差因素的干扰,而基于误差传播定律的常规方法对各类误差的处理策略使得海底点坐标解算不准确。针对这一问题,本文将声速测距误差非时变项设为待解参数,在水下观测方程的系数矩阵中讨论声速测距误差时变项与换能器位置误差的影响,构建了GNSS-声学水下定位的动态非线性高斯-赫尔默特(Gauss-Helmert, GH)模型,并推导了该模型的总体卡尔曼滤波解。在此基础上,进一步考虑扩展后的观测信息受到粗差污染的情况,给出了模型的抗差处理方法及解算步骤。最后分别通过仿真试验和胶州湾海域实测试验进行了验证,试验结果表明,在不同深度或不同换能器位置误差大小的无粗差设定下,本文方法解算精度及稳定性较常规方法均更高;当观测信息含有粗差时,模型的抗差滤波算法能更准确地识别及定位异常信息,其三维点位精度明显更优,解算效果达到最佳。

The GNSS-acoustic combined observing is an important means to determine the position of seafloor control points,but it will be interfered by error factors such as the uncertainty in sound velocity and the positioning deviation of the sea surface platform.However,the processing strategy of general method based on the error propagation law for various errors makes the seafloor point coordinate solution inaccurate.To solve the above problems,this paper sets the time-invariant term of sound velocity ranging as the parameter to be solved,and discusses the influence of time-varying error of sound velocity ranging and transducer position error in the coefficient matrix of underwater observation equation.Thus,the dynamic nonlinear Gauss-Helmert(GH)model for GNSS-acoustic underwater positioning is constructed,and the total Kalman filter solution of this method is derived.On this basis,taking into account the gross errors polluting of the observations,the robust method and solution steps of the new model are given.Finally,simulation experiments and a testing experiment in the sea area near Jiaozhou Bay are used to verify the performance of the new model.The results show that under conditions with no gross errors and either different water depths or different transducer position errors,the accuracy and stability of the proposed method are both higher than those of the general method.When the observations are polluted by gross errors,the robust filter algorithm of the new model can accurately identify and locate the abnormal information.The precision of its 3D point deviation results can be obviously optimized,and the solution performance is superior to that of the general method.