STMicroelectronics MEMS microphones lay the foundation for future voice-activated home development

     STMicroelectronics will play an important role in the European Project Distant Speech Interaction for Robust Home Applications (DIRHA). This three-year project aims to investigate and propose solutions to the problems that natural speech human-computer interaction will face in future smart home applications.

     DIRHA is designed to address the challenges of long-distance speech human-computer interaction in noisy, multi-speaker environments. The goal of the project is to create an environment where voice information can be received anytime, anywhere, and recognized and used by the system even if the user is not speaking into an authorized microphone. From an acoustic point of view, the speech recognition system must extend to the speaker's side, and the speech recognition performance is not affected by the speaker's location in the home.

     The visual dimensions and acoustic parameters of ST's MEMS microphones fully meet the stringent requirements of remote speech interaction systems. The compact size of MEMS microphones makes it easy for researchers to embed an entire array of microphones fully into the walls, desks, or voice-activated appliances of an automated home, while the superior acoustic properties of the microphones, combined with advanced signal processing technology, make it easier to integrate the microphones into an automated home. Enables a voice system to recognize and capture a person's voice commands from a few meters away in a room with many people playing music.

     Long-distance voice human-computer interaction can not only dramatically change the way humans interact with machines, but it can also truly change the lifestyle of people with limited mobility, such as the elderly or those with movement disorders. In addition to home use, remote voice human-computer interaction systems are also suitable for robotics, telepresence, security surveillance and industrial automation.

     The DIRHA project is divided into a number of working groups, with a total duration of 36 months and a total investment of 4.8 million euros. His main research areas include multichannel acoustic processing, long-distance speech recognition, identification/verification, and spoken conversation management (German, Greek, Italian, and Portuguese). The final design prototype will be deployed in a trial home, where human users will evaluate the design's performance.