一种新的整体最小二乘迭代算法

otal Least Squares problem was first proposed by Golub and Van Loan.They have given the first numerically stable solution.After more than twenty years of research,people have given the sufficient condition and the form of solution of solvable Total Least Squares problem.In recent years,there are more practical solution of Total Least Squares problem such as SVD method and iterative method.This paper first introduce existing common solution of Total Least Squares problem,pointing out the defects of these methods when used in practical.Furthermore,we propose a new iterative algorithm for Total Least Squares problem based on these methods.Data measurement field has been plagued by precision evaluation of Total Least Squares problem,this paper presents strategy of precision evaluation and generalize and define the degree of freedom of Total Least Squares problem.Finally,we verify the feasibility of this new algorithm through project example.