| 摘要 |
In an environment of temporally and spatially uncorrelated white additive noise, the direction estimation apparatus for coherent signals calculates (M-1) instantaneous cross-correlations between some array data, and selecting multiple pairs each of which is consisted of p instantaneous correlations from the (M-1) instantaneous cross-correlations to form a Hankel correlation matrix. Next, the apparatus divides that Hankel matrix into an upper submatrix and a lower submatrix and then calculates a linear operator at time n by using the adaptive LMS algorithm with a fixed or time-varying step-size parameter and or using the adaptive NLMS algorithm from the two submatrices. Then a noise subspace is estimated from that linear operator. Finally the directions of incident signals at time n is estimated and tracked by using that noise subspace with Newton approximation.
|
