SPC06-5: Turbo Equalization with Integrated SNR Estimation

We propose a turbo receiver structure with integrated SNR estimation for frequency selective channels. The proposed receiver is able to estimate the signal-to-noise ratio (SNR) and improve the estimation via iteration. Initially we exploit a subspace-based method to estimate the SNR. During the iterative process of equalization and decoding if the quality of the a posteriori information at the output of decoder improves, it is possible to utilize this information as an alternative to the original subspace method to mitigate any mismatch between the actual SNR and the estimated SNR at the receiver. A hypothesis testing criterion is introduced to decide whether or not this transition is beneficial in terms of the receiver performance. Simulation results demonstrate that our proposed receiver structure is able to yield a performance similar to that of the known SNR at the receiver.