Stability Analysis of Nonlinear Digital Filters with External Interferences

Abstract

Filtering, a class of signal processing serves as an elementary operation in various applications. The fundamental problems in the practical design of a recursive digital filter are the nonlinear effects arising from finite wordlength nonlinearities and external interference. The finite wordlength nonlinearities are the composite nonlinearities originating from several combinations of quantization and overflow. Due to these finite wordlength effects, the designed recursive digital filter may tend to oscillate and becomes unstable. On another research front, ringing which is the resultant of undesired effects due to past excitations also leads to oscillations. Considering the presence of external interference, improper operations and/or failures may result in the implemented digital filter. So, it is necessary to have sophisticated algorithms or criteria that facilitate the designer in choosing filter specifications so that the designed digital filter becomes free of oscillations. The main intent of this work is to find suitable criteria under which the designed digital filter is asymptotically stable with enhanced performance against external interference. The demand for improved algorithms and enhanced devices is an increasing process in the field of technology since, Today s luxury is tomorrow s necessity.