AI project to monitor Covid-19 patients with ultrasound

October 26, 2020 //By Nick Flaherty
AI project to monitor Covid-19 patients with ultrasound
Ten Fraunhofer Institutes in Germany are working on an internal project to develop a modular, distributed AI-based monitoring system for Covid-19 patients using a MEMS ultrasound sensor.

The ten Fraunhofer groups are linking with four university hospitals and clinics in German on the M³Infekt project that will enable early intervention if a patient’s condition suddenly starts to deteriorate. The project is based around a MEMS ultrasonic transducer to monitor breathing and includes open standard interfaces, sensor fusion and machine learning.

One of the challenges with Covid-19 that has become apparent during the pandemic is that patients can exhibit mill symptoms and then go downhill very quickly. The mobile system being developed by the M3Infekt cluster project will acquire, analyse and fuse relevant biosignals to monitor a patient.

As part of the M³Infekt project, Fraunhofer’s IPMS Photonic Microsystem Institute will  work on a spirometric breath analysis using a 128-channel CMUT MEMS-based ultrasound device. The capacitive MEMS ultrasound array enables the monitoring of the airflow of the exhaled breath and so monitors the breathing of the patient in general. It is also developing a system for chemical breath analysis based on ion mobility spectroscopy.

The Frauhofer MEOS project hub will coordinate the clinical partners, work on the regulatory issues for measurements and manage the system integration.

The aim is to provide a long-term solution for decentralized monitoring of patients on normal wards and in non-hospital environments monitoring heart rate, ECG, oxygen saturation, blood flow as well as respiratory rate/volume and breath analysis. Machine learning methods serve as the basis for evaluating these parameters, supporting the diagnosis and enabling integration of the system into different deployment and application scenarios, regardless of location.

The planned system has a modular and mobile structure with standardized, open interfaces. These enable easy integration into other platforms and make the system suitable for use with various diseases, including influenza, pneumonia and sepsis.

This will enable continuous monitoring, previously used only for patients in intensive care, to be rolled out to non-hospital scenarios, such as short- and long-term care, outpatient treatment or home settings. This way,


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.