Multirate digital systems have important applications in communication and signal processing including resampling, interpolation, narrow band filtering, fax systems, subband coding, transmultiplexing, multiresolution analysis, and waveletts. This research is developing algorithms for the systematic design of multirate signal processing systems based on a model-matching approach where an ideal system (model) is approximated by a practical multirate system consisting of components such as digital filters, modulators, and upsamplers/downsamplers. The problem is to choose the parameters of the practical system so that the performance of the system is as close as possible to the ideal system. The measurement of the difference between systems used in this research is based on the difference in outputs of the two systems for all possible inputs. System parameters are chosen to minimize the maximum energy of the difference between output signals for all bounded energy input signals. This error measure gives systems where the quality of the approximation is insensitive to the particular type of input to the system. The most significant improvements over conventional designs are expected in complicated systems where intuitive design methods are limited, such as in the design of cascaded multirate systems.
