Angle Estimation for Adaptive Linear Array using PCA-GS-ML Estimator

The maximum likelihood (ML) angle estimator can yield optimal angle estimation performance. In the work presented here, a fast algorithm for solving the global optimal solution of the ML angle estimator based on principal component analysis (PCA) and grid search (GS) is developed. Utilizing the low-rank property of the mainbeam steering matrix, the log-likelihood function can be decomposed as a combination of the relevant quantities of basis vectors of the low-rank subspace. Thus, evaluation of the log-likelihood function can be realized in a lower dimensional space. Although GS is also required, the computational complexity can be greatly reduced, and the global optimal solution can be obtained.