A Speech to Machine Interface Based on Perceptual Linear Prediction and Classification
Saeed Mian Qaisar, Noofa Hainmad, Raviha Khan, Rawan Asfour · 2019 Advances in Science and Engineering Technology International Conferences (ASET) · 2019
The recent technological advancements are focusing on developing the smart systems to facilitate the subscribers and to improve their lifestyle. The machine learning algorithms and artificial intelligence are becoming the elementary tools, which are used i n the establishment of smart cities and smart buildings across the globe. In this context, automatic speech recognition based system is devised. It can be effectively integrated in smart spaces like buildings, offices, homes, etc. The system digitally processes the acquired speech command and extracts its Perceptual Linear Predictive Coding Coefficients (PLPCC). IN next step a comparison of the extracted parameters is made with the templates in order to recognize the incoming command. The command recognition is achieved with a specifically developed voting based classifier. The classification algorithm is described. A method to characterize the developed classifier is also discussed. On the detection of a command the embedded controller is piloted by a specific flag. The front-end controller decodes this flag and performs a desired action by piloting a specific actuator. A system prototype is realized. Preliminary, its application is demonstrated as a speech driven curtain controller. The front end controller is implemented with an Arduino board and motors are employed as curtain drivers. The prototype functionality is successfully tested and results are presented.