A MIMO parameter estimation model taking Ricean fading channel and stochastically uncorrelated signals into account. Part II. Asymptotically efficient estimators

Abstract

The purpose of this paper is twofold; to demonstrate parameter estimation possibility of MIMO model and to propose an asymptotically efficient approach for estimating three channel parameters such as, nominal direction, angular spread and rice factor. Manipulating the estimation in efficient manner, the benchmark and the proposed estimators are based on weighted least squares (WLS) criteria. Rather than invoking the sample covariance estimate in ordinary WLS criterion, the proposed approach makes use of structured covariance estimate for computing the weight matrix. The nonparametric estimate presented here in is the way to constrain the array covariance matrix to hold the Toeplitz structure so that the residual due to imposing the structured weight matrix is less than that provided by employing the ordinary sample covariance. As a matter of course, this leads to more advantage for parameter estimation, particularly, in non-asymptotic region of temporal snapshot number. Numerical examples are conducted to validate the superiority in non-asymptotic situations.