Maximum likelihood estimation of 2-D superimposed exponentialsignals

The problem of maximum likelihood estimation of the parameters (i.e. frequencies, amplitudes, and noise variance) of 2-D superimposed exponential signals is considered. In this paper, a polynomial rooting approach is proposed to obtain the estimates of the 2-D frequencies. Strong consistency and limiting distribution are established for the estimates of the parameters. Furthermore, the covariance matrix of the limiting distribution attains the Cramer-Rao lower bound