From: DOA estimation based on data level Multistage Nested Wiener Filter
Forward recursion for i = 1, 2, ⋯, D | |
\( {\mathbf{t}}_i={\displaystyle {\sum}_{n=1}^{L-1}{d}_{i-1}^{*}(n){\mathbf{x}}_{i-1}(n)} \) | |
t i  = t i /‖t i ‖ 2 | |
\( {d}_i(n)={\mathbf{t}}_i^{\mathrm{H}}{\mathbf{x}}_{i-1}(n),\kern0.24em n=1,\cdots, L-1 \) | |
x i (n) = x i− 1(n) − d i (n)t i , n = 1, ⋯, L − 1 | |
ε D (n) = d D (n), n = 1, ⋯, L − 1 | |
Backward recursion for i = D − 1, ⋯ , 2, 1 | |
\( {\omega}_{i+1}={\displaystyle {\sum}_{n=1}^{L-1}{d}_i(n){\varepsilon}_{i+11}^{*}\left[n\right]/}{\displaystyle {\sum}_{n=1}^{L-1}{\left|{\varepsilon}_{i+1}(n)\right|}^2} \) | |
ε i (n) = d i (n) − ω i  + 1 ε i+ 1(n), n = 1, ⋯, L − 1 | |
Calculate the Wiener filter coefficient | |
\( {\mathbf{w}}^{(D)}={\displaystyle {\sum}_{i=1}^D{\left(-1\right)}^{i+1}\left\{{\displaystyle {\prod}_{l=1}^i}\left(-{\omega}_l\right)\right\}\times {\mathbf{t}}_i} \) |