A CNN Based Speech Recognition Approach for Voice Controlled Elevator

Disabled people face many challenges in day-to-day life. One of the challenges is to use a lift in large buildings. Being unable to use a lift, taking stairs has increased the number of accidents and injuries. To avoid such scenarios, the proposed system acts as an intermediate between humans and machines. Convolutional Neural Network (CNN) based speech recognition method used in the proposed system recognizes the voice command given by the person. The Mel Frequency Cepstral Coefficient (MFCC) features are used. Passengers can give voice commands in Marathi regional language and English language. The system will be useful to communicate with ease and also contactless. Performance of proposed algorithm verified on Marathi, English and both voice commands. This experimentation gave an accuracy of 95%, 85%, and 75% on Marathi, Englishand both language voice commands respectively. The lift is fully automated and there is no need for human intervention which provides all the disabled people with ease of life.