Enhanced adaptive filtering using second order statistics

In the area of adaptive filter there is a compromise between convergence performance and computational complexity. The LMS adaptive algorithm has simple computational load, but on the other hand has poor convergence properties. There are methods used to enhance its convergence rate like the transform domain adaptive filter (TDAF). In this paper the use of second order statistics (SOS) as a preprocessing technique is proposed. Simulation results illustrate successful performance of signal detection at low signal to noise ratio. This proposed method improves the filter convergence rate and stability.