Many high resolution methods such as MUSIC rely on a signal subspace decomposition, which is conventionally accomplished by an eigendecomposition. However an eigendecomposi-tion is very computationally intensive and difficult to implement. In this paper, by using the special structure of the covariance matrix, a fast high resolution method based on QR decomposition is presented. In most case, the performance of the new method is similar to those of MUSIC method, but the new method has smaller computational complexity and easier for parallel implementation.
Schmidt R O. IEEE Trans. on AP, 1986, AP-34(3): 276-280.[2]Kumareson R, Tufts D W. IEEE Trans. on. AES, 1983, AES-19(1): 154-159.[3][3][4]Kaveh M, Barabell A J. IEEE Trans. on. ASSP, 1986, ASSP-34(2): 331-341.[5]Roy R H, Kailath T. IEEE Trans. on. ASSP, 1989, ASSP-37(7): 981-995.[6]Prasad S, Chandna B. IEEE Trans. on SP, 1991, SP-39(5): 1224-1228.[7]李有明,保铮.电子学报,1994, 22(2): 47-52.[8]李有明.电子科学学刊,1994, 16(6): 641-645.[9]Gill P E, et al. Math. Comp., 1974, 28(2): 505-535.
DownLoad: