Skip to main content

Time Encoded Voice Interfaces

Project description

Sensor encoding and interactions

Ultra-low power sensors incorporating artificial intelligence require an analog-to-digital converter (ADC) before the signal processing block implemented with a neural network (NN). ADCs have been employed based on a voltage-controlled oscillator (VCO) to implement integrated sensors. This requires the development of a new multiply–accumulate cell (MAC) for the first layer of the NN that operates with signals from the VCO as well as a suitable VCO interface with existing sensors. However, methods aiming to apply a sensor interaction using a VCO and implement a phase/frequency-encoded MAC unit for the NN have been attempted separately. The EU-funded TEVI project aims to combine the two for the benefit of research and industry applications. To that end, it will remove the ADC before the NN by directly connecting the sensor to it and encoding the sensor signals with a VCO.

Coordinator

UNIVERSIDAD CARLOS III DE MADRID
Net EU contribution
€ 752 714,64
Address
Calle Madrid 126
28903 Getafe (Madrid)
Spain

See on map

Region
Comunidad de Madrid Comunidad de Madrid Madrid
Activity type
Higher or Secondary Education Establishments
Non-EU contribution
€ 0,00

Participants (1)

INFINEON TECHNOLOGIES AUSTRIA AG
Austria
Net EU contribution
€ 0,00
Address
Siemensstrasse 2
9500 Villach

See on map

Region
Südösterreich Kärnten Klagenfurt-Villach
Activity type
Private for-profit entities (excluding Higher or Secondary Education Establishments)
Non-EU contribution
€ 0,00