A modified maximum likelihood method for SNR estimation in OFDM based systems

In this paper, a modified Maximum Likelihood (ML) method is proposed to estimate the noise variance as well as the corresponding signal to noise ratio (SNR) in Orthogonal Frequency Division Multiplexing (OFDM) based receivers under Additive White Gaussian Noise (AWGN) as well as multipath fading channels without using any pilot symbols. The presented algorithm works directly in the time domain (pre-FFT method) by using the redundant information of the cyclic prefix and can be applied to any unitary or non-unitary constellation. Furthermore, all extracted parameters can be used to evaluate the signal quality within the receiver and build up applications such as Minimum Mean Square Error (MMSE) channel estimation or adaptive synchronization algorithm design, especially for Software Defined Radio (SDR) applications. All simulations are applied to the Digital Audio Broadcasting (DAB/DAB+) system and show better performance compared to the state of the art SNR estimation procedures.