A Novel TDLMS Algorithm Based on the Fractional Fourier Transform

Gang Su,Guangxi Zhu
DOI: https://doi.org/10.1109/WiCOM.2006.195
2007-01-01
Abstract:LMS adaptive filtering is widely used in smart antenna and space-time signal processing. The transform domain LMS (TDLMS) adaptive filtering is a very popular alternative technique to general LMS-type algorithms. As Fourier analysis is one of the most frequently used tools in signal processing, this paper first introduces fractional Fourier transform (FRFT) and part of its properties, and then presents its interpretation as a rotation of the time-frequency plane. The FRFT's relationships with time-frequency representations have a very simple and natural form and supports LMS adaptive time-frequency filtering in the FRFT domain. The simulation results indicate that the novel method can speed convergence effectively compared to those existing transform domain approaches in literature
What problem does this paper attempt to address?