Kaya, IsmailTugcu, EminÖzen, AliNix, Andrew Robert2025-02-242025-02-242015978-1-4799-8088-82577-24652-s2.0-84940417232https://hdl.handle.net/20.500.14440/76981st IEEE Vehicular Technology Conference (VTC Spring) -- MAY 11-14 -- 2015 -- Univ Strathclyde -- Technol Innovat Ctr -- Glasgow -- SCOTLANDA novel fast blind equalizer is obtained by using the direct calculations from a channel matched filter decision feedback equalizer (CMF-DFE). The proposed technique converts the inverse convolution operations of an equalizer into a linear finite impulse response estimation filter, which is more suitable for blind training. A novel error function is introduced for blind training which enables the use of fast algorithms such as LMS or RLS. The required auto-regression values for the CMF-DFE equalizer are calculated from the incoming data. The resulting performance with LMS training is close to that of non-blind techniques.eninfo:eu-repo/semantics/closedAccessSelf-Recovering EqualizationConstant Modulus AlgorithmTrackingConference ObjectN/AWOS:000371404700254N/A