Decomposition of a linear-filter channel with additive noise

This paper introduces a new multiplexing system for discrete-time signals based on polynomial residue arithmetic. Using this multiplexing system, a linear additive-noise channel is decomposed into independent subchannels that are also modeled as quasi-linear additive noise channels. This property is useful for subchannel users to apply accumulated conventional technologies in the design of subchannel communication systems, and to satisfy their specific needs independently. The proposed multiplexing system also utilizes the whole transmission bandwidth efficiently. For proper selection of polynomials, subchannel information is dispersed uniformly in both frequency and time so that contributions of frequency or time localized degradation average over the subchannels. The averaging has an effect of decreasing the total error probability of transmission for such distortion channels. Analysis exhibits this improvement.