Abstract:

To improve the estimation performance and decrease the calculation complexity, a direction of arrival estimation of coherent signals named ToepUESRPIT is proposed. The estimation of signal parameters via rotational invariance techniques(ESPRIT) instead of multiple signal classification(MUSIC) is used to avoid the peak searching. Based on the Toeplitz matrix, centroHermitian matrices are constructed twice. Also, eigenvalue decomposition and total least square problems are changed into real ones by unitary transformation to reduce the computational complexity. Besides, the algorithm essentially doubles the number of available measurements. Increased estimation accuracy can, therefore, be achieved. Simulation results and analysis prove the validity and effectiveness of the algorithm.