Skip to Main Content
The current channel estimation algorithm is often performed under the assumption that the additive channel noise is white and Gaussian, but there is experimental evidence to show that the additive channel noise is non-Gaussian and non-linear in a wireless environment. This paper addresses the MIMO-OFDM channel estimation based on particle filtering and under the Middleton Class A noise model. The proposed algorithm models the wireless fading channel as an AR process, and optimizes the particles distribution using a variable step-size gradient information. Compared with conventional estimation approaches, the proposed method outperforms in robust to non-Gauss distribution noise. The simulations show the effectiveness of the new scheme.