SNR Estimation for Preamble-based Wireless OFDM Systems using Extended Kalman Filter

Abstract: In Orthogonal Frequency Division Multiplexing (OFDM) systems, robustness in frequency selective channels is achieved using adaptable transmission parameters. To reckon these parameters, knowledge of Signal to Noise Ratio (SNR) estimates obtained by channel state information is essential. This necessitates for an appropriate channel estimation scheme to acquire efficient SNR estimates in wireless frequency selective fading channels. Improved Periodic Sequence (IPS) based OFDM system incurs SNR estimates by utilizing Least Squares (LS) channel estimates and adaptively choosing significant Channel Impulse Response (CIR) paths in Discrete Fourier Transform (DFT) interpolation. LS channel estimation scheme is a linear processing method, which disposes for only linear characteristics of wireless channels. In order to contend with the non linearity of frequency selective wireless channels, a non linear Extended Kalman Filter (EKF) estimation scheme is implemented with DFT interpolation in this extended IPS estimation algorithm. The proposed extended IPS estimator outperforms IPS estimator in terms of average SNR and SNR per subcarrier for frequency selective channels.