MULTICHANNEL FAST QRD-LS ADAPTIVE FILTERING - NEW TECHNIQUE AND ALGORITHMS

In this paper, a direct, unified approach for deriving fast multichann el QR decompostion (QRD) least squares (LS) adaptive algorithms is int roduced. The starting point of the new methodology is the efficient up date of the Cholesky factor of the input data correlation matrix. Usin g the new technique, two novel fast multichannel algorithms are develo ped. Both algorithms comprise scalar operations only and are based exc lusively on numerically robust orthogonal Givens rotations. The first algorithm assumes channels of equal orders and processes them all simu ltaneously, It is highly modular and provides enhanced pipelinability, with no increase in computational complexity, when compared with othe r algorithms of the same category. The second multichannel algorithm d eals with the general case of channels with different number of delay elements and processes each channel separately. A modification of the algorithm leads to a scheme that can be implemented on a very regular systolic architecture. Moreover, both schemers offer substantially red uced computational complexity compared not only with the first algorit hm but also with previously derived multichannel fast QRD schemes. Exp erimental results in two specific application setups as well as simula tions in a finite, precision environment are also included.