Radio Sensing and Semantic Sensing

Next-generation communication networks enable much more than just data transmission and positioning. The same radio signals that are used today for networking and location tracking can also serve as sensors: they monitor the status of equipment, detect movements and anomalies in production processes, and support safety and security functions – without additional sensor hardware.

At Fraunhofer IIS, we research and develop solutions for radio sensing and semantic sensing as central building blocks for the digital transformation of industry, logistics, and safety-critical applications.

Radio Sensing: Funksignale als Sensorik

Radio sensing: radio signals as sensors

Radio technologies such as 6G, 5G, UWB, WLAN and BLE already form the digital nerve pathways of companies today. These networks continuously generate radio signals whose propagation is influenced by objects, people and movements in the environment.

By systematically evaluating various channel parameters, information about the environment can be derived, for example:

• Presence of people and objects
• Movement in certain zones

This creates an invisible, comprehensive sensor layer that builds on existing radio infrastructure without the need to install additional hardware.

Semantic sensing: from signal to meaning

While radio sensing detects physical changes in the radio field, semantic sensing goes one step further: it combines radio signal and environmental data with AI methods and machine learning to derive semantic information. Semantic sensing thus enables context-aware interpretation of radio data and supports:

• Environment classification (e.g. material recognition)
• Process automation (e.g. detection of disruptions in the process)
• Anomaly detection (e.g. early detection of drones)

Typical challenges:

• Downtime is only detected at a late stage.
• Material shortages are only noticed after the fact.
• Process deviations lead to rejects or delays.
• Traditional sensor technology (cameras, tags, light barriers) is expensive, maintenance-intensive or critical in terms of data protection.

With radio sensing and semantic sensing, manufacturing processes can be analysed continuously and non-invasively, for example:

• Detection of downtime and blockages at stations
• Identification of deviating processing times or cycle disturbances
• Monitoring of material flow and occupancy status

This opens up new opportunities for:

• Process optimisation
• Early detection of problems (predictive process monitoring)
• Lean retrofit solutions for digitising existing facilities without major conversion

Examples include:

• Entering secure areas: When a person enters a defined security area, the signals from several radio links change in a characteristic way. These patterns can be automatically evaluated and recognised as an area violation.
• Monitoring of airspace: If flying objects (e.g. drones) violate the airspace above an area requiring protection, this can also be detected by the change in the radio field.

Sensing monitors entire volumes and can be integrated into existing communication infrastructures. Sensing offers advantages over traditional systems in this regard:

• Cameras are not permitted or can only be used to a limited extent for data protection reasons.
• Light barriers and light curtains usually only monitor individual lines.
• Radar systems are complex systems that ideally require a clear line of sight.

With our technologies and projects in the field of radio and semantic sensing, we demonstrate how existing radio networks can be transformed into powerful sensor and localisation platforms for industry, logistics, and safety and security.

Reconfigurable Intelligent Surfaces (RIS)

With reconfigurable intelligent surfaces (RIS), we are investigating how radio waves can be specifically shaped to revolutionise connectivity while significantly increasing the performance of sensing and 6G localisation systems.

DeltaPro

In the DeltaPro project, we are researching and demonstrating how radio networks can be used for precise positioning and radio sensing in real production environments to make processes more transparent, efficient and secure.

XGDensify

In the XGDensify project, we are developing an end-to-end pipeline for environmental perception: from 3D models and material recognition to ray tracing simulations, in order to create realistic models for 5G and future 6G networks.