A pseudo-multirate implementation of digital systems for improved performance

The digital implementation of systems implies some accuracy degradation due to the combined effect of sampling and finite wordlength, and said effect is particularly especially evident if the system dynamics are spread in a wide frequency range. This paper discusses an implementation scheme based on a multirate discretization approach that may improve the representation accuracy of the digital system as opposed to a standard single-rate approach. A framework is also introduced for defining feasibility conditions in terms of accuracy requirements. The proposed multirate discretization approach can be implemented on standard single-rate hardware devices, whence its qualification as "pseudo-multirate". Some simulation results are provided to illustrate the effectiveness of the proposed methodology.