| （英） |
Digital filters play an important role in a wide range of signal (media) processing areas, including speech processing, image processing, and digital communications. In particular, a variable digital filter can be regarded as a typical reconfigurable signal processing system that is able to perform online characteristic tuning. More specifically, a variable digital filter can update its coefficients in real time and regenerate a new frequency-domain characteristic in a fast manner as needed. However, stability-guarantee is a central and unavoidable issue in designing a variable recursive filter. This is because a variable recursive filter changes its coefficients on the fly during signal processing, which may cause instability. Basically, designing a variable recursive filter belongs to a nonlinear programming problem, and such a design formulation without stability constraints are most likely to produce a unstable recursive filter. Up to the present, the author has proposed an effective measure based on parameter-transformations to ensure the stability. This report aims to reconfirm the rationale for the stability-constrained design formulation through computer simulations. The computer simulations verify that it is most unlikely to get a stable recursive filter if the design is done without imposing any stability constraints. In other words, the computer simulations are highly consistent with the well-proved stability theories, and the stability-guarantee measures using parameter-transformations work effectively. |