Correlation-based blind SINR estimation with statistical theoretic analysis for OFDM systems

Abstract A blind signal-to-interference-plus-noise ratio (SINR) estimation method is proposed for orthogonal frequency division multiplex (OFDM) systems in the presence of impulse noise over multipath fading channels. According to the multipath channel order and length of cyclic prefix (CP), the received signal on each OFDM symbol is divided into six data intervals which reveal that there is inter-symbol interference (ISI) in the 1st and 6th intervals and no ISI in the 2nd to 5th intervals. The data in the 1st, 2nd, and 3rd intervals, respectively, are replicated data in the 4th, 5th, and 6th intervals. Based on these observations, first, the multipath channel order can be estimated by the autocorrelation functions of the data in the 1st, 2nd, 4th, and 5th intervals based on likelihood function. Second, utilizing the properties that there is no ISI and autocorrelation functions of noise plus interference are zero in the 2nd data interval, the signal powers can be estimated from the correlation coefficients of the received signals in this interval. Third, since the CP data on the 2nd interval are a copy of the data in the 5th interval, the interference plus noise power can be obtained based on the autocorrelation function of the data difference between these two intervals. Based on the above analysis, a closed-form expression of SINR estimation is obtained. Furthermore, closed-form expressions for the normalized mean square errors (NMSE) and the Cramer–Rao lower bound (CRLB) for the proposed SINR estimation are also derived. Both computer simulation and theoretical analysis results show that the proposed method achieves a lower NMSE than the cyclic correlation-based method and NMSE of the proposed method is closer to the CRLB.