Server-based Speech Technologies for Mobile Robotic Applications

Paper proposes the server-based technologies and the overall solution of the multimodal interface (speech and touchscreen) usable for mobileapplications in robotics as well as in other domain. The server-based automatic speech recognition server, able to handle several audio input streams, has been designed, developed and connected to the Android application. It receives input data stream and sends back the recognition result. The second important technology was designed and implemented to synthesize artificial speech. The server-based TTSsolution was prepared and connected. HMM-based approach was applied including recording and training of new voices. Finally, the simple client application for Android devices was developed and tested. Thediscussion of related problems is proposed also in the paper