Author(s): Abdullah M.K. | F.I.Shaikh | A.I.Tamboli
Journal: International Journal of Computer Applications
ISSN 0975-8887
Volume: iccia;
Issue: 5;
Date: 2012;
Original page
Keywords: Adaptive filters | affine combination | convex combination | least mean square (LMS)
ABSTRACT
This paper studies the statistical behavior of an affine combination of the outputs of two least mean-square (LMS) adaptive filters that simultaneously adapt using the same white Gaussian inputs. The purpose to combine two filters is to obtain a new LMS adaptive filter with fast convergence and small steady-state mean-square deviation (MSD).The linear combination studied is generalization of convex combination in which combination factor ?(n) is restricted to the interval (0,1).Each of the two filters produces dependent estimates of unknown channel. Thus there exists a sequence of optimal affine combining coefficients which minimizes the mean-square error (MSE) and gives good steady state response. First optimal unrealizable affine combiner is studied. Then two schemes proposed to find out the optimal mixing parameter to get optimized sequence ?(n) are stochastic gradient approach and error power based scheme. The mean square performances are analyzed and validated by MATLAB7.
Journal: International Journal of Computer Applications
ISSN 0975-8887
Volume: iccia;
Issue: 5;
Date: 2012;
Original page
Keywords: Adaptive filters | affine combination | convex combination | least mean square (LMS)
ABSTRACT
This paper studies the statistical behavior of an affine combination of the outputs of two least mean-square (LMS) adaptive filters that simultaneously adapt using the same white Gaussian inputs. The purpose to combine two filters is to obtain a new LMS adaptive filter with fast convergence and small steady-state mean-square deviation (MSD).The linear combination studied is generalization of convex combination in which combination factor ?(n) is restricted to the interval (0,1).Each of the two filters produces dependent estimates of unknown channel. Thus there exists a sequence of optimal affine combining coefficients which minimizes the mean-square error (MSE) and gives good steady state response. First optimal unrealizable affine combiner is studied. Then two schemes proposed to find out the optimal mixing parameter to get optimized sequence ?(n) are stochastic gradient approach and error power based scheme. The mean square performances are analyzed and validated by MATLAB7.