整体最小二乘可以有效解决变量误差（EIV）模型的求参问题，受到了测绘、信号处理等众多领域的关注。然而，受限于EIV模型的标准形式，相应的整体最小二乘算法在一些非线性问题的研究中无法严格适用，一定程度上限制了其应用推广。为此，本课题拟以广义EIV或非线性高斯-赫尔默特(GH)模型为基础研究广义整体最小二乘理论，并拓展其粗差处理、方差分量估计和推估算法，具体包括：（1）基于敏感度分析建立等式约束条件下非线性GH模型的数据探测算法，用于处理各种受粗差影响的等式约束最小二乘和整体最小二乘问题；（2）提出非线性GH模型的抗差LS-VCE算法，抑制非线性模型中的观测粗差的同时合理调整各类观测值之间的权比；（3）提出约束广义整体最小二乘推估算法，提高约束整体最小二乘问题的预测精度。研究成果有助于完善整体最小二乘的理论和应用体系，实现测量基准转换、精密工程测量、水文大地测量等相关领域中的高精度参数估计。

Total least squares (TLS) can effectively solve the issue of parameter determination of EIV model, and thus has been paid attentions in many areas, including geomatics and signal processing etc. However, due to the standard form limitations of the EIV model, the corresponding TLS algorithms are not rigorously applicable in the studies of some nonlinear problems, which restricts the applications and popularizations of the TLS to some extent. For this reason, the theories of generalized TLS (GTLS) are investigated based on the generalized EIV or nonlinear Gauss-Helmert (GH) model in this project, and the corresponding outliers processing, variance component estimation (VCE), and prediction algorithms are extended, including: (1) a data snooping algorithm for equality constrained nonlinear GH model is established based on the sensitivity analysis, which aims to deal with all kinds of constrained LS and TLS problems influenced by outliers; (2) the robust LS-VCE algorithm for the nonlinear GH model is proposed, so as to simultaneously resist gross errors and adjust the relative weight ratios of different types of observations in nonlinear models; (3) a constrained generalized TLS prediction algorithm is proposed, which aims to improve the prediction accuracy of the constrained TLS problems. The research outputs of this project will help to perfect the theory and application systems of TLS, and consequently achieve the high precision parameter estimation in the fields of surveying datum transformation, precise engineering surveying, and hydrogeodesy etc.