Information processing in passive elastic structures
Smart devices constantly monitor the environment and analyse the resulting stream of information, waiting for wake-up events ranging from a skipped heartbeat in a pacemaker to a verbal command in an intelligent speaker. In electro...
ver más
¿Tienes un proyecto y buscas un partner? Gracias a nuestro motor inteligente podemos recomendarte los mejores socios y ponerte en contacto con ellos. Te lo explicamos en este video
Proyectos interesantes
TEC2011-28250-C02-02
SISTEMAS BIO-INSPIRADOS DE RADIO COGNITIVA BASADOS EN ELECTR...
63K€
Cerrado
PID2020-120271RB-I00
ACELERADORES BASADOS EN FPGAS PARA REDES NEURONALES PROFUNDA...
77K€
Cerrado
TEC2015-66878-C3-1-R
INTEGRACION DE COMPONENTES Y ARQUITECTURAS PARA LA VISION EM...
198K€
Cerrado
TEC2010-21196-C02-01
SISTEMAS AUTONOMOS TOLERANTES A FALLOS BASADOS EN DISPOSITIV...
8K€
Cerrado
TIN2008-06856-C05-05
TECNOLOGIAS DE FUSION SENSORIAL AUDIO-VISUAL PARA SISTEMAS D...
27K€
Cerrado
PID2019-105556GB-C32
SENSORES CONTROLADOS POR EVENTOS PARA APLICACIONES CON INSPI...
84K€
Cerrado
Información proyecto INFOPASS
Duración del proyecto: 64 meses
Fecha Inicio: 2022-01-31
Fecha Fin: 2027-05-31
Descripción del proyecto
Smart devices constantly monitor the environment and analyse the resulting stream of information, waiting for wake-up events ranging from a skipped heartbeat in a pacemaker to a verbal command in an intelligent speaker. In electronic devices, such constant workload requires reliable power and leads to significant battery drainage. The aim of INFOPASS is to answer the need for zero standby power, always-on sensing and processing, by inventing novel components based on architected elastic microstructures. We will focus on passive speech recognition: Vibrating structures that respond with large amplitudes only when excited by a particular spoken word. The components will be based on artificial neural networks of mechanical resonators -- benefiting from their ultra-low power dissipation and bypassing the inefficient transduction between acoustic and electric signals. The goal of passive speech recognition is challenging due to the required device complexity. Conventional micromechanical systems consist of a few resonators, while neural networks require thousands of nodes, pushing the limits of design and fabrication technologies by orders of magnitude. The proposed work builds on my invention of a two-step structure design method: Advanced responses are engineered by first encoding the desired response in an effective mass-spring model, which is then translated into a structural geometry. Compared to full-wave simulations, mass-spring models speed-up the optimisation process by more than a thousand-fold, allowing us to 'train' the structures on realistic speech corpuses. The invention of devices that passively perform complex information processing tasks will lead to mobile devices with longer battery life, always-on Internet-Of-Things devices that consume no standby power, and battery-less health monitoring sensors ? potentially impacting billions of devices.