A highly accurate and low bias SNR estimator: Algorithm and implementation

Signal to noise ratio (SNR) estimators are required for many radio engineering applications. In this paper, a SNR estimator based on the first and second order moments is derived and examined for constant envelope modulations over additive white Gaussian noise (AWGN) channel. Firstly, a modification method is proposed to reduce the bias of conventional first and second order moments based SNR estimator. Then, in order to reduce hardware implementation complexity, multi segments of cubic polynomial are utilized to approximate the nonlinear inverse function of the proposed SNR estimator. The approximate expression is very precise in the SNR range from -10dB to 20 dB as illustrated by numerical simulation results. Besides, practical hardware circuit is proposed for FPGA implementation. Simulation results show that the proposed SNR estimator has the lower normalized bias and variance when compared with two other classic estimators.