Maximum-likelihood based estimation of angular parameters of Dense-Multipath-Components

This contribution introduces an estimation framework for angular/spatial properties of the Dense Multipath Components (DMC). The estimator aims to calculate the parameters of an antenna-independent DMC model from the (antenna-dependent) spatial sample covariance matrix. A Newton-Raphson based iterative algorithm with “Fisher Scoring” is used to solve the maximum-likelihood problem and a least-mean-squares method is proposed to initialise the polarisation parameters of the model. The feasibility of the approach is shown using simulations that incorporate measured antenna array radiation pattern.