Czyz K.

The paper presents a new design concept of a compact active noise control (ANC) system employing unique signal sampling and reconstruction methods. The classical approach to the ANC system design assumes necessity of high order analogue filters to avoid aliasing effects. The main drawback of this approach is that these analogue filters complicate dynamics of the controlled plant. This makes control process itself much more difficult as well as raises the final cost of the controller. The different possible techniques replacing analogue filters in the system design were considered. Nonuniform signal sampling occurred to be the most effective. The deliberate irregularities in signal sampling, can suppress aliasing without introducing any additional dynamics into the control path. Moreover, in the presented system additional oversampling method replaced classical signal reconstruction basing on analogue forming filters. The inter-sample effects in reconstructed signal are still present but moved beyond audible frequency range. The presented concept allowed to develop ANC system that has efficient hardware implementation with simple analogue frontend. It enables fast reconfiguration of the sampling frequency according to application needs with no hardware modifications. System employing nonuniform sampling of the error and the reference signals with oversampling of the control signal was used for creation of local spatial zones of quiet in enclosures. The results of real-world experiments confirmed that the elimination of anti-aliasing filters leads to higher convergence speed and greater noise attenuation level in the ANC system, especially for random disturbances.