期刊名称：International Journal of Advances in Engineering and Management
电子版ISSN：2395-5252
出版年度：2020
卷号：2
期号：4
页码：504-509
DOI：10.35629/5252-0204480485
语种：English
出版社：IJAEM JOURNAL
摘要：Voice has become one of the effective and efficient communication medium among people. This communication method may also be a useful interface for interacting with machines which understand only electrical signals. Thus, the system 's reliance on speech recognition has increased dramatically over the last year. Various methods of speech recognition exist, such as Dynamic Time Warping (DTW), Vector Quantization (VQ),Hidden Markov Model (HMM), etc. This paper offers an detailed analysis over the use of Artificial Neural Networks ( ANN)in speech recognition. Here we used a different class of neural networks called Multilayer Perceptron (MLP), which uses the back propagation for error algorithm. Once data has been collected, the voice signal is pre-processed and fed to the MLP for classification. This paper introduces the prototype of speech recognition-based automation system to control the various electrical devices for the differently abled people suffering from quadriplegia or paraplegia (who can not move their body parts, but can talk and listen). Using the MATLAB (2018a) environment to monitor electrical equipment, the automatic speech recognition (ASR) system is used to accurately and effectively transform a voice signal into a text message.
