This paper presents the concept and formulation of a finite-memory filter for real-time estimation of noise-corrupted signals. It is assumed that an autoregressive moving average (ARMA) model of the dynamic process is available. The structure of the filter is nonrecursive, and is based on weighted averaging of a finite array of past values of measured inputs and outputs of the dynamic process. The performance of the proposed filter has been evaluated by simulation experiments.
|Original language||English (US)|
|Number of pages||5|
|Journal||American Society of Mechanical Engineers (Paper)|
|State||Published - 1992|
All Science Journal Classification (ASJC) codes
- Mechanical Engineering