| Abstract [eng] |
Presented work covers a noise reduction system development (noise cancellation headphones). The functionality of the system is based on active noise cancellation. Main objective is to develop noise reduction filter, which is based on fuzzy inference systems and artificial neural networks functionality. The analysis part of presented work goes through fuzzy inference systems and artificial neural networks analysis, recurrent fuzzy neural filter analysis and selection of most suitable recurrent fuzzy neural filter according to analysis results. The dynamic fuzzy neural network is selected as noise reduction filter. The structural identification technique of selected noise reduction filter is developed. The presented technique represents a systematic approach to generating fuzzy rules from a given input-output data set. The dynamic fuzzy neural network was implemented and simulated using MatLab program package. The experimental research was made using real measurements from laboratory of acoustics to evaluate the results when system is functioning in real environment. |
