Time and frequency selective deconvolution uising optimal control

A recently developed approach to deconvolution uses the methods of optimal control. The technique has been successfully applied to a number of problems including geophysics and speech processing. In this paper we demonstrate that the optimal control method can provide complete control over the deconvolution in both time and frequency domains, and that in consequence it is a powerful technique for deconvolving noisy and severely bandlimited signals. In the first instance it is shown that through the parameters we can control the resolution of the deconvolution in the time domain. Secondly it is demonstrated that the effects of inaccurate estimation of the system parameters may be approximately described by the ratio of the transfer function of the true system to the estimated. From this the bandwidth of the resulting deconvolution can be arbitrarily controlled by mismatching the system estimates in an appropriate manner.