Example 2 Suppose is a zero-mean random variable which is to be estimated from two zero-mean random variables and . Let the LMMSE estimator be . Then the optimal values of and are given by

           This results in the orthogonality conditions .

and    

           Rewriting the above equations, we get