
Decentralized patient monitoring: Sensors quickly detect changes in vital signs
M³Infekt is a one-year cluster project conducted by a consortium of several Fraunhofer Institutes and clinical partners headed by the Fraunhofer Institute for Integrated Circuits IIS. The project which was launched in September 2020 aims “to develop a multi-modal, modular and mobile system of sensors for monitoring infectious diseases,” explains Dr rer. nat. Christian Münzenmayer, Head of Department, Image Processing and Medical Engineering at Fraunhofer Institute for Integrated Circuits IIS.
Interview: Sascha Keutel
What exactly are the objectives the Fraunhofer Institute wishes to achieve with M³Infekt?
Currently, vital signs such as ECG, blood pressure, oxygen saturation, perfusion or breath gas are monitored only in the ICU but not on the general ward. During the Corona pandemic, however, we have seen many cases with mild symptoms deteriorating rapidly. Thus we suggest mobile monitoring of vital signs so the patients can remain on the general ward as long as possible while the care staff can intervene quickly when the patient’s status worsens.
Does that mean the monitoring system is designed for containing SARS-CoV2?
The project M³Infekt is funded under the action programme “Fraunhofer vs. Corona”. Under this programme, the Fraunhofer-Gesellschaft invests millions of euros in the development of technologies to conquer the pandemic and to boost the economy.
Thus Covid-19 is simply an example for a current field of application. With the modular and mobile architecture of the sensors we want to develop a system that can be used for several diseases, for example pneumonia and influenza. In general, the system should be suitable for all diseases that require continuous monitoring.

What kinds of technologies are being used?
We are planning to develop separate sensor components which will lead to a full system, including a sensor bracelet that captures biosignals such as heartbeat and oxygen saturation.
Moreover, we will be using a textile-integrated multichannel ECG system that our team at Fraunhofer IIS originally developed for sports and wellness purposes. This device, called CardioTEXTIL, is a holster with integrated ECG electrodes and a small electronics unit that transmits data to the docking station or the smartphone. We want to adapt this system for clinical monitoring – we are thinking about a portable device, a wearable monitor for the patient.
We will also be using radar-based and hyperspectral sensors for contact-less recording of biosignals. These optical technologies use different wavelengths in order to be able to measure heartrate or breathing frequency.
Imagine a system that is mounted above the hospital bed and transmits data via radio signals. The idea behind this concept is to keep patients on the ward mobile rather than tying them down with cables.
Recommended article

Breath gas analysis detects cancer
Researchers are developing solutions designed to enable the analysis of breath gas to assist with the diagnosis of disease.
For breath gas analysis we are looking into two different technologies: an MEMS-based ultrasound system that can monitor breath gas flow. MEMS means Micro-Electro-Mechanical Systems – tiny integrated components that link microelectronic and mechanics. We are not talking about a conventional PC chip but about microsystems with integrated mechanics.
In addition we are looking into chemical breath gas analysis with ion mobility spectrometry. This allows simple and quick capturing of relevant volatile organic compounds, so-called VOCs. VOCs are present in exhaled breath, urine, stool and skin and can be measured and analysed on-site.
The team is also working on a sound measurement technology that analyses the spatial lung ventilation, reconstructs it in 3D and thus allows measuring it over a certain period of time.
We want to create a wide portfolio of different sensor technologies. Some of these technologies are new developments, others already exist and we adapt them for our purposes. This includes algorithms that interpret the enormous data volumes to detect any change in the patient’s health status.
Is the system designed exclusively for clinical use?
Initially, the product is meant for clinical or paraclinical use such as care facilities. By the end of the project term we hope to have developed individual modules that can perform first measurements and can be used in trials.
At that point, there won’t be a finished and market-ready product yet. We rather aim to develop a concept with standardized open interfaces which allow easy integration with other platforms. And we expect different devices to be developed in the context of the M³Infekt cluster project.
A long-term aim might well be a system that can be uses in home monitoring or out-patient settings. Therefore, we will design follow-up projects for which we will be looking for industry partners who can support commercialisation efforts.

You mentioned algorithms: are they AI processes?
We are working at different algorithms on different levels. On the one hand there is on-sensor signal processing which can be based on an embedded chip in a wearable device. This technology could be used for evaluating ECG signals. A microcontroller with AI procedures based on neural networks will be implemented to determine ECG parameters.
This technology has a major advantage: raw signals do not have to be transmitted to a docking station and thus less bandwidth is used. This is also a privacy issue since ECG data is like finger prints – it can be traced back to individual patients. We can overcome this privacy challenge by pre-evaluating data in the device.
On the other hand we are aggregating and analysing different health parameters provided by the individual sensors to analyse the overall health status. In short: from ECG data we determine heart rate or heart rate variability as a parameter, or from breath gas flow we might be able to determine maximum inhaling or exhaling depth.
In a final step we combine all these parameters in order to be able to assess the patient’s health status and its development. This is based primarily on AI procedures but also on conventional signal processing.
How will the data be prepared for the medical staff?
The staff will see the data on a dashboard in the physicians’ or nurses’ room. This will enable them to assess the crucial parameters right away and it will offer early indicators in case the patient’s health status deteriorates.

Profile:
Dr Christian Münzenmayer studied informatics at the University of Applied Sciences Augsburg and computational engineering at Friedrich Alexander University Erlangen-Nürnberg. He received his doctoral degree from the University Koblenz-Landau with a dissertation on machine learning in medical applications. He moreover holds an MBA from FOM University of Applied Sciences for Economics & Management, Essen.
In 2000, he joined the Fraunhofer Institute for Integrated Circuits IIS in Erlangen. In his current role as Head of Department, Image Processing and Medical Engineering, he is in charge of the division Digital Health Systems at Fraunhofer IIS. He is also a member of the management team of the division Smart Sensing and Electronics.
You are a professional from the emtech ecosystem and interested in sharing your tech tale with a guest contribution at Tectales? Just contact us: editorial@tectales.com.