15.02.17

MIT chip offers huge power savings for voice technology

hand holding smartphone using voice recognition software

Huge players like Amazon and Google are not the only ones backing voice technology to continue to impact our everyday lives, as MIT has recently announced a low-powered chip designed for automatic speech recognition.

Researchers at Massachusetts Institute of Technology (MIT) are backing voice-controlled devices to become more integrated with how we live, and has introduced a chip offering power savings between 90-90% compared to normal mobile phones running speech recognition software.

For example, an iPhone running Siri may use 1W of power whereas the chip developed by MIT could do same work for .2 to 10 milliwatts (depending on what’s required of it) – a huge development for the use of mobile and wearable devices, as well as power-constrained devices in the wider IoT.

The chip is able to conserve energy via a "voice activity detection" circuit that monitors ambient noise, and only activates its more complex, energy-heavy speech recognises circuit when it detects sounds that actually constitute speech.

It is also able to compress data and evaluate a small section of audio within its onboard memory to keep energy usage low.

"Speech input will become a natural interface for many wearable applications and intelligent devices," commented Anantha Chandrakasan, professor of Electrical Engineering and Computer Science at MIT, whose group developed the new chip.

He continued: "The miniaturisation of these devices will require a different interface than touch or keyboard. It will be critical to embed the speech functionality locally to save system energy consumption compared to performing this operation in the cloud."

"For the next generation of mobile and wearable devices, it is crucial to enable speech recognition at ultralow power consumption," added Marian Verhelst, professor of microelectronics at the Catholic University of Leuven, Belgium. "This is because there is a clear trend toward smaller-form-factor devices... Speech offers a very natural way to interface with such devices."

The team from MIT presented their paper detailing the chip at the Solid-State Circuits Society conference earlier this month.

Source: MIT