Spoken Dialogue-based Remote Control of Educational Mobile Robot with Mecanum Wheels

Abstract

This paper describes a new platform of educational mobile robot with Mecanum wheels and voice enabled dialogue-based human robot interaction model was established. Moreover, new trends in using deep neural networks for cloud- and edge-based speech recognition was described in detail. Finally, the progress of both the cloudand edge-based speech recognition was reported. Especially, promising usability test result on resisting room reverberation and background noise had been achieved by the proposed convolutional neural network (CNN)-based wake-up word (WUW) spotting system.