The design of sum-of-cisoids channel simulators using the iterative nonlinear least square approximation method

Abstract

In this paper, we propose the iterative nonlinear least square approximation (INLSA) algorithm as an effective method for the design of sum-of-cisoids (SOC) channel simulators assuming non-isotropic scattering conditions. For the characterization of non-isotropic scattering scenarios, we use the von Mises distribution for describing the distribution of the angles-of-arrival (AOAs). The INLSA method relies partially on numerical optimization techniques. This method determines the SOC model parameters iteratively by minimizing the Frobenius error norm. We evaluate the performance of the INLSA method and compare the results with those obtained for the Riemann sum method (RSM) and the Lp-norm method (LPNM). The performance comparisons will be carried out with respect to the autocorrelation function (ACF) and the distribution of the envelope. The obtained results indicate that the proposed method is more preferable than the RSM and the LPNM when emulating the statistical properties of the channel under non-isotropic scattering conditions. In addition to its efficiency and excellent performance, the simplicity of the INLSA makes the method a powerful tool for designing channel simulators required for the test and evaluation of mobile communication systems.