Multichannel Blind Deconvolution Using the Conjugate Gradient

In this paper, we propose a conjugate gradient based algorithm for blind deconvolution. In general, blind deconvolution algorithms suffer from the speed of convergence. We make a further study of the geometrical structures on the manifold of finite impulse response (FIR) filters using lie group method. We derive the expressions of geodesic and parallel translation on the manifold of FIR filters. Using mutual information criteria. a feasible cost function is derived for blind deconvolution problem. Then we develop a conjugate gradient algorithm for multichannel blind deconvolution problem in finite impulse response (FIR) manifold. Computer simulations show the validity and effectiveness of this approach.