摘要
基于矩阵的SVD理论 ,通过修改设计阵的奇异值 ,提出了若干种直接解算病态 (满秩或秩亏 )观测方程的方法 ,以改善设计阵的病态性 ,提高参数估值的稳定性和准确性。
Based on Singular Value Decomposition (SVD), several methods for directly solving ill conditioned equations with full rank or not full rank are proposed by modifying the singular values of the design matrix. The numerical results show that the proposed methods can resist the ill conditioned effect and improve the numerical stability and accuracy of parameter estimation.
出处
《大地测量与地球动力学》
CSCD
2004年第3期15-18,共4页
Journal of Geodesy and Geodynamics
基金
国家杰出青年科学基金项目 (4 0 12 5 0 13 )
国家自然科学基金项目 (4 0 0 740 0 6)
关键词
平差模型
SVD理论
观测方程
奇异值分解
解算方法
surveying equation, ill-conditioned, singular-value decomposition,adjustment model,solving method
