最小二乘及其改进算法在外测数据处理中的应用

摘要

在外测数据处理中,经典最小二乘法的使用经常出现局限性,通过引进Jacobi矩阵,讨论了多台雷达定轨的非线性加权最小二乘估计问题;对于带有先验信息的测量数据,采用Bayes估计解算飞行轨迹;为解决测量数据的设计矩阵多重相关性问题,采用有偏的岭估计公式进行轨迹推算.对于不同外测方案和各种误差特性的测量数据,应根据数据客观情况选择适宜的算法,为最小二乘及其改进算法在外测数据处理中的应用提供新思路.%The classical Least Squares algorithm is limited in measured data post-processing. This paper discusses the ques-tion of least square estimate for multi-radar nonlinear model by importing Jacobi matrix, adopts Bayes estimate for the meas-ured data with prior information, uses ridge estimation formula to resolve design matrix multiple correlations of measured da-ta. For different test project and measured data with all kinds of errors, fit formula should be chosen bases the factof meas-ured data. This paper provides a new way to the application of Least Squares and its improved algorithm in measured data post-processing.