An intelligent control system for positioning a ventilator using acoustic feedback

A knowledge-based control strategy has been implemented to control the position of a ventilator for the purpose of noise control. The level of intrusive transportation noise into the building is controlled by activating an intelligent controller using acoustic feedback when the noise occurs. There are three major steps of training, source classification and control in which fuzzy logic and rule-based algorithms are implemented. If somebody plays music or speaks near the window, the intelligent window must not react even though the level of the sound may be higher than an approaching aircraft or vehicle. To identify an approaching noise source, estimated parameters of a second order autoregressive model AR(2) and the peaks of short time Fourier transform STFT are used. The cross-correlation between output data from indoor and outdoor microphones indicates the effect of outdoor transportation noise on the indoor noise level. If the dominant indoor noise is transportation noise its level is used as acoustic feedback in a rule-based control loop. A training software is employed to update the thresholds and to generate the triangle membership functions with three linguistic variables for decision making and control scenario. The experimental results of application a prototype intelligent window is also presented.