Application of IoT using neuro-fuzzy based on thai speech classification to control model hospital bed with arduino

Abstract

© 2019 IEEE. Nowadays, the number of chronic patients who had to stay in the hospital is high. The bedridden patients need carers to help them in many activities such as giving medicines, eating food, flipping their body. The development of prototype in this research, including software and equipment relieves the burden for carers or helps the bedridden patients. The software is a mobile application controlling hardware by touch screen or using Thai voice. The hardware is a model hospital bed controlling by an Arduino. The hardware and software are connected via Wi-Fi. To implement Thai voice command in the mobile application, the Neuro-fuzzy with Mel frequency Cepstral coefficients is selected to create the Thai speech classification model. There are several experiments to find the best structure. The average accuracy of 5-fold cross-validation of the best model in testing data is 71.50%.