Title :
Multichannel linear and quadratic adaptive filtering based on the Chandrasekhar fast algorithm
Author :
Sayadi, Mounir ; Fnaiech, Farhat ; Najim, Mohamed
Author_Institution :
Ecole Superieure des Sci. et Tech. de Tunis, Tunisia
fDate :
3/1/1999 12:00:00 AM
Abstract :
A new fast algorithm for multichannel linear and quadratic adaptive filtering using the Chandrasekhar equations is presented. Based on the shift-invariance property, the multichannel linear model could be described by a time-invariant state-space model to which we apply the Chandrasekhar factorization technique, which provides interesting numerical properties. Furthermore, a new method for nonlinear filtering is given where the multichannel Chandrasekhar algorithm is applied on the second-order Volterra (SOV) filter after suitable transformations
Keywords :
Volterra series; adaptive Kalman filters; adaptive signal processing; filtering theory; nonlinear filters; parameter estimation; state-space methods; Chandrasekhar equations; Chandrasekhar factorization; Chandrasekhar fast algorithm; Kalman filter; multichannel linear adaptive filtering; multichannel linear model; nonlinear filtering; numerical properties; parameter estimation; quadratic adaptive filtering; second-order Volterra filter; shift-invariance property; time-invariant state-space model; truncated Volterra series; Adaptive algorithm; Adaptive filters; Filtering algorithms; Finite impulse response filter; Maximum likelihood detection; Nonlinear equations; Nonlinear filters; Polynomials; Riccati equations; Signal processing algorithms;
Journal_Title :
Signal Processing, IEEE Transactions on
