Principle of least squares (3) - recursive least squares

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Because the   covariance matrix, only the elements on the diagonal there, so it does not affect the transposition, i.e.

 

In fact, recursive least squares: just wanted to minimize the sum of the variance, and then found that the variance is the sum of the estimated tracking error covariance matrix, which also contains the trace K, K so that it is seeking the smallest trace, that is K derivation.

 

Reference link: https: //zhuanlan.zhihu.com/p/59532437