Compressed and Deep Sensing: Models, Structures, and Tasks

Informazioni relative al progetto

CoDeS

ID dell’accordo di sovvenzione: 101000967

DOI

10.3030/101000967

Data della firma CE 30 Aprile 2021

Data di avvio 1 Luglio 2021

Data di completamento 30 Giugno 2026

Finanziato da

• EXCELLENT SCIENCE - European Research Council (ERC)

Costo totale

€ 3 000 000,00

Contributo UE

€ 3 000 000,00

Coordinato da

WEIZMANN INSTITUTE OF SCIENCE
Israel

Periodic Reporting for period 2 - CoDeS (Compressed and Deep Sensing: Models, Structures, and Tasks)

Periodo di rendicontazione: 2023-01-01 al 2024-06-30

Sintesi del contesto e degli obiettivi generali del progetto

In this project we are exploring how we can build systems that are far more efficient than those existing today by designing them in an end-to-end holistic fashion. Throughout their design we take into account ay structure in the input or system, and we also consider the final system task. We have been exploring ways in which this approach can make various types of systems more efficient including communication systems, radar, medical devices and more. In all of these areas we consider the basic question: Can we measure only what is necessary to perform our desired system task by jointly optimizing the entire signal acquisition and processing chain?

Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti

In our work so far we have been developing fundamental trade-offs between sampling rate, quantization resolution, and task fidelity. We have also developed new results on the bounds of model-based deep learning that takes models into account within the learning process.
On the engineering side we have developed several new ADC devise based on this principle like modulo-based sampling that can accommodate high dynamic range, and time-based sampling that allows removing the system clock all together. On the application side we have been considering 6G networks, joint sensing and communication systems, efficient vital sign monitoring, ECG monitoring and more.

Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora)

On the engineering side we have developed several new ADC devise based on this principle like modulo-based sampling that can accommodate high dynamic range, and time-based sampling that allows removing the system clock all together. On the application side we have been considering 6G networks, joint sensing and communication systems, efficient vital sign monitoring, ECG monitoring and more.