Hardware-optimized mobile biosignal analysis

AI-based analysis of biosignals from wearables

Particularly in the recording of biosignals from wearables, latent motion artifacts or noise often make reliable data analysis difficult. The risk of error is exacerbated by a lack of computing capacity sufficient for recording and processing the vast quantities of data.

We at Fraunhofer IIS develop AI-based algorithms for analyzing biosignals from mobile measurement systems. These can be implemented in an energy-efficient way on various hardware platforms and allow for reliable data interpretation, even when motion artifacts are present.

By optimizing neural networks, we can provide near-sensor AI implementation with real‑time data evaluation.

Examples of available algorithms   

  • Analysis of respiratory parameters
  • Sensor data fusion
  • Analysis and reconstruction of human motion (e.g. walking, running, cycling)
  • Detection of heart rate and heart rate variability
  • Arrhythmia detection (e.g. atrial fibrillation)

If you want to get more than just random noise out of your data set, give us a call – we would be happy to find the right solution for you!

We license our existing algorithms or develop individual analysis solutions tailored to our customers’ requirements.

Your advantages with us as a partner

Biosignal analysis in real time and under everyday conditions

  • Evaluation of discreet, mobile measurement systems in which the algorithms used can produce robust results, even on the basis of low-quality input
  • Optimization of deep neural networks (DNN) for near-sensor, energy-efficient analysis
  • Multimodal data evaluation by combining multiple sensors to record biosignals (e.g. respiration and heart rate)
  • Comprehensive, interdisciplinary network and alliances with universities, hospitals and medical engineering companies as well as sound domain knowledge of biosignal analysis
  • Fully equipped facilities with test laboratories and a high-performance computing cluster

Please note

Our working methods are modeled closely on the CE marking process applicable in Europe; certification based on regulatory medical requirements will not be conducted.

More Information


Communication. Sensors. Analysis.

Digital Health Systems


Each document, such as application note, brochure or others, and each video, can be found in our download section.

Development Partner for Personalized Medical Engineering

Research and development services from a single source