The paper provides the high dimensional optimal problem from which the RELAX algorithm is derived, and compares the RELAX algorithm with the alternating projection algorithm from some aspects. Based on the discussion, an improved RELAX algorithm is proposed. The computer simulation confirms the theory is validity.
Li J, Stoica P, Zheng D. Angle and waveform estimation in the presence of colored noise via RELAX. submitted to IEEE Trans. on Aerospace and Electronics Systems,1995.[2]Li J, Stoica P. Efficient mixed-spectrum estimation with applications to target feature extraction.[3]IEEE Trans. on Signal Processing. 1996, SP-44(2): 281-295.[4]邵朝,保铮.RELAX算法的相关域分析.西安电子科技大学学报,1997, 24(2): 164-171.[5]Stoica P, Sharman K, Maximum likelihood. Methods for direction-of-arrival estimation. IEEE Trans Acoustics, Speech and Signal Processing, 1990, ASSP-38(7): 1132-1143.[6]Ziskind I, Wax M. Maximum likelihood localization of multiple sources by alternating projection. IEEE Trans. on Acoustics, Speech and Signal Processing. 1988, ASSP-36(10): 1553-1560.[7]Hwang J K, Chen Y C. Superresolution frequency estimation by alternating notch periodgram. IEEE Trans. on Signal Processing, 1993, SP-41(2): 727-741.[8]Viberg M, Ottersten B. Sensor array processing based on subspace fitting. IEEE Trans. on AP, 1991, AP-39(5): 1110-1121.[9]Bresler Y, Macovski A. Exact maximum likelihood parameter estimation of superimposed exponential signals in noise. IEEE Trans. on Signal Processing. 1993, SP-41(2): 727-741.[10]Golub G H, pereyra V. The differentiation of pseudo-inverses and nonlinear least squares problems[11]whose variables separate. SIAM J[J].Numer. Anal.1973, 10(2):413-432