贝叶斯重力平差算法的稳定性评估与统计检验

Stability evaluation and statistical test of Bayesian gravity adjustment algorithm

采用Bootstrap和Jackknife两种统计学方法,用于评估贝叶斯重力平差算法的参数估计稳定性问题.以四川和云南实际重力测网的观测系统为例,模拟观测数据和误差分布,给出了贝叶斯重力平差方法用于实际重力平差的稳定性评估方案.研究结果表明:贝叶斯平差方法在观测样本足够的情况下,可以较好地估计相对重力仪的非线性漂移和格值系数;部分段差样本丢失时仍可以稳定地估计模型参数.通过使用统计学方法和检验流程,并结合特定的重力观测系统进行测试,所得结论有助于优化陆地重力观测系统设计,评估贝叶斯重力平差结果的可靠性.

Bootstrap and Jackknife statistical methods were used to evaluate the stability of the parameter estimation of the Bayesian gravity adjustment algorithm.Taking the observation systems of Sichuan and Yunnan actual gravity survey networks as examples,the stability evaluation scheme of Bayesian gravity adjustment method applied to actual gravity adjustment was given by simulating the observation data and error distribution.The results show that the Bayesian gravity adjustment method can accurately estimate the nonlinear drift of the relative gravimeter and the scale factor when the observation samples are enough.The model parameters can be estimated stably even when part of the step samples are lost.The statistical methods and testing procedures used in this study can be tested in combination with specific gravity observation systems,and the conclusions obtained are helpful to optimize the design of terrestrial gravity observation systems and evaluate the reliability of Bayesian gravity adjustment results.