Subspace tracking based on the constrained projection approximation approach

In this paper, we present an algorithm for tracking the signal subspace recursively. It is based on an interpretation of the signal subspace as the solution of a minimization of a constrained projection approximation task. We show that we can apply the matrix inversion lemma to solve this problem recursively. Proposed algorithm avoids orthonormalization process after each update for post-processing algorithms which need orthonormal basis of the signal subspace. Simulation results in the direction of arrival (DOA) tracking context depict high performance of this algorithm in comparison with other algorithms.