| description abstract | A narrow frequency bandwidth, strong fluctuations of the gain versus signal frequency and sensitivity to disturbances caused by the operating environments are the most common factors limiting the applicability of sensors in manufacturing systems. Self-tuning filters represent an efficient means of alleviating these limitations. Since the dynamic properties of sensors vary rapidly, a successful implementation of sensors coupled with self-tuning filters hinges upon accurate, real-time adjustments of these filters. The selection of optimum filter settings, based upon the available distorted output signals from the in-process sensors, poses a difficult problem. In general, the algorithm of self tuning requires a priori information about the sensor and its environment, condensed into a form of an analytical model. A systematic approach to the analytical modeling of sensors is proposed. To illustrate this approach, a comprehensive model of a commercial dynamometer is developed and tested. | |
