时间相关观测的几种动态数据处理方法

现代大地测量数据中往往存在时间相关性，忽略观测值时间相关性会影响参数估值的精度和可靠性。因此，本文研究了几种时间相关观测的动态数据处理方法。首先，对观测值时间相关性的处理方法进行了扩展和统一。利用极大验后估计原理，提出了第3种思路，并在无历元公用参数和有历元公用参数两类模型下，详细推导了去相关变换法、差分变换法和极大验后估计法这3种思路的动态解，并对它们的特点和等价性进行了论证。其次，为了平衡实际应用中的解算效率且同时有效顾及时间相关性，利用自相关函数表达时间相关性，导出了相应退化形式。最后，通过GPS实测数据验证了本文理论公式的正确性和实用性。

Several Kinematic Data Processing Methods for Time-correlated Observations

Time correlations always exist in modern geodetic data, and ignoring these time correlations will affect the precision and reliability of solutions. In this paper, several kinematic data processing methods for time-correlated observations are studied. Firstly, the method for processing the time-correlated observations is expanded and unified. Based on the theory of maximum a posteriori estimation, the third idea is proposed. Two types of situations with and without common parameters are both investigated by using the decorrelation transformation, differential transformation and maximum a posteriori estimation solutions. Besides, the characteristics and equivalence of above three methods are studied. Secondly, in order to balance the computational efficiency in real applications and meantime effectively capture the time correlations, the corresponding reduced forms based on the autocorrelation function are deduced. Finally, with GPS real data, the correctness and practicability of derived formulae are evaluated.