Towards Laplacian angle deviation model for spatially distributed source localization

The paper aims at presenting a more realistic Rayleigh channel model for the problem of finding the arrival direction in the presence of a spatially spread source. Unlike the previous angular deviations assumed to be randomly distributed as uniform or normal (Gaussian), a Laplacian assumption is taken into account instead of those so as to support recent field measurements. With regard to the inherent accuracy imposed in our realistic model, we provide a rather systematic formulation of Cramer-Rao bound at hand and a derivative expression of the Laplacian characteristic function aligned in the angular spread parameterization matrix. Incorporating a maximum likelihood estimator into the model, some numerical results demonstrate that there exist small and considerable degradations of nominal angle and angular spread estimates owing to the heavier tail of the Laplacian PDF.