Blind Receiver Design for OFDM Systems Over Doubly Selective Channels
We develop blind data detectors for orthogonal frequency-division multiplexing (OFDM) systems over doubly selective channels by exploiting both frequency-domain and time-domain correlations of the received signal. We thus derive two blind data detectors: a time-domain data detector and a frequency-domain data detector. We also contribute a reduced complexity, suboptimal version of a time-domain data detector that performs robustly when the normalized Doppler rate is less than 3%. Our frequency-domain data detector and suboptimal time-domain data detector both result in integer least-squares (LS) problems. We propose the use of the V-BLAST detector and the sphere decoder. The time-domain data detector is not limited to the Doppler rates less than 3%, but cannot be posed as an integer LS problem. Our solution is to develop an iterative algorithm that starts from the suboptimal time-domain data detector output. We also propose channel estimation and prediction algorithms using a polynomial expansion model, and these estimators work with data detectors (decision-directed mode) to reduce the complexity. The estimators for the channel statistics and the noise variance are derived using the likelihood function for the data. Our blind data detectors are fairly robust against the parameter mismatch.
© Copyright 2007 IEEE. Reprinted with permission. Paper approved by Y. Li, the Editor for Wireless Communication Theory of the IEEE Communications Society. Manuscript received April 29, 2005; revised July 23, 2006. [Posted online: 2007-05-15] This work was supported in part by the Natural Sciences and Engineering Research Council of Canada, the Informatics Circle of Research Excellence, and the Alberta Ingenuity Fund. This paper was presented in part at the IEEE International Conference on Communications, May 2005, Seoul, Korea.