Novel blind channel amplitude and phase spectra estimation technique using 2nd order statistics

Blind channel estimation using 2^nd order statistics of the over-sampled received signal with the exploitation of signal cyclo stationary properties, has gained much attention recently. A novel estimation procedure for a base-band communication channel parameters using a frequency domain model is proposed in this paper. The described method provides useful insight to the channel estimation problem. The frequency domain approach works by independently and blindly estimating the channel amplitude and phase spectra. This approach has also been extended for evaluating estimation bounds (CRB $Cramer-Rao bounds) of channel parameters. Furthermore, the proposed frequency domain technique does not depend on the noise probability density function and could be easily extended for evaluating the estimation performance in the presence of colored noise.