Automatic environmental noise source classification model using fuzzy logic Burak Uzkent, Buket D. Barkana ⇑ , Jidong Yang School of Engineering, University of Bridgeport, 221 University Ave., Bridgeport, CT 06604, USA article info Keywords: Fuzzy logic ACF feature parameters Environmental noise recognition abstract Noise classification is very important nowadays. Fuzzy logic has been applied to many interesting prob- lems in different areas including noise identification/recognition. With this study, we propose an auto- matic environmental noise source classifier based on fuzzy logic. The proposed classifier uses the feature parameters that are extracted using short-time auto-correlation function. Six commonly encoun- tered non-stationary noise sources are chosen to recognize. These are subway, airport, inside car, inside train, restaurant, and rain. Classification accuracy of the proposed classifier ranged from 62% to 90% rates. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Automatic noise identification and classification has become a very active subject of research during last decades, since it can be directly or indirectly implemented into a very wide area of top- ics including speech recognition, pattern recognition, and context aware applications. A large amount of work on environmental noise classification and identification has been completed to sepa- rate speech from background noise for robust speech recognition; however a few context-aware applications have attempted to use environmental noise sources. In noise monitoring systems, classifi- cation of environmental noises has been provided to help in con- trolling noise pollution (Couvreur & Bresler, 1995). Adaptive information systems are another research area using environmen- tal noise classification, since the environmental noise can provide a rich source of information about the current context. Different methods and algorithms are developed for identifying noise sources (Abdou & Guy, 1996; Kenji, Masatsugu, & Yoichi, 2004; Lan & Chiu, 2008; Noël, Planeau, & Habault, 2006; Yang, Wang, Hao, Shen, & Qi, 2009; Zhang, Gan, & Zhou, 2006). While one group of studies focuses on the extraction of the feature parameters, the other group of studies focuses on the classification techniques such as Hidden Markov model (HMM), statistical pat- tern recognition systems, neural networks, and fuzzy logic systems. During the past few years, as the complexity of issues increases, it is difficult to design an appropriate system to achieve a sharp classification goal. In the consideration of this matter, a degree of membership in the fuzzy logic became a new method of solving the problems. Fuzzy classification creates more subtle and smooth distinctions between equivalence classes than traditional crisp classification (Kaufmann & Meier, 2009). Although the fuzzy logic is relatively young theory, the system based on fuzzy rules has been broadly used in range of problem domains to solve classifica- tion problems. In the literature (Puzzolo, De Natale, & Giannetti, 2003), a fuzzy classification approach of multi-temporal SPOT (HRV) data was tested in order to increase forest type discrimina- tion at the specific level in an alpine mountain area. Fuzzy theory also can be used as Fuzzy indexer for ancient document and corre- sponding typesets (Sousa, Gil, & Pinto, 2007). There is a wide use of fuzzy logic techniques to model classifi- cation of speech, image, and background noise in signal processing area. Chen and Wang proposed an efficient noise reduction method by using fuzzy logic approach (Chen & Wang, 2009). They stated that their proposed fuzzy logic approach achieves a fairly better performance than a number of other existing methods. Another fuzzy logic approach for classification of background noise (Beritel- li, Casale, & Ruggeri, 1999) proposed a background noise classifier based on fuzzy logic approach and it is compared with Quadratic Gaussian Classifier (QGC). They seperated the inputs according to parameter of being stationary and nonstationary. They report that fuzzy logic approach reduces misclassification between stationary and nonstationary background noise while increasing the classifi- cation accuracy %10 more than Quadratic Gaussian Classifier. An- other study is reported on acoustic noise identification using fuzzy logic techniques (Silva, Sousa, Botto, & Sá Da Costa, 2007). They propose fuzzy identification as a nonlinear tool to derive acoustic noise models. According to their results, fuzzy logic is more accurate than linear models and the model outputs are com- puted using less computational effort. Mostly, the manmade and natural noises in nature are non-sta- tionary, however most methods studying noise sources have been developed for the stationary situations. To understand the differ- ence between stationary and non-stationary noise relays on the nature of the signals. Stationary noise sources have a relatively 0957-4174/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2011.01.084 ⇑ Corresponding author. E-mail address: bbarkana@bridgeport.edu (B.D. Barkana). Expert Systems with Applications 38 (2011) 8751–8755 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa