The goal of this project is to design and develop a system, which enables the waiter of a restaurant to take orders by voice. For that, the waiter needs to pronounce the order to his PDA, and the integrated voice recognition engine will translate the pronounced order to text. Afterward, the text is submitted to the restaurant’s management software solution for further processing.
One general problem of voice recognition based products is their sensitivity to noise. To address this problem, the HES-SO VS has developed a dedicated hardware device, which is plugged onto the PDA.
Its task consists in extracting the waiter’s voice from the recorded audio signal in order to provide a high quality signal to the recognition engine. This processing is very important for noisy environments (such as restaurants) and has been realized under the guidance of EPFL’s acoustic laboratory.
Besides that, a novel speech recognition engine developed by IDIAP has been integrated into the system and the necessary Middleware - responsible for the data exchange between PDA, voice recognition engine and the dedicated device - has been written.
This project has been financially supported by the Swiss Innovation Promotion Agency.