Intelligent home presence, without compromising privacy
WINP is a research and development project focused on the study and application of WiFi sensing technology to produce an accessible, low-cost real-time indoor locationsystem.
The project explores an innovative approach that uses existing WiFi networks not only for connectivity, but as a tool for perceiving the environment, overcoming the limitations of GPS technologies and traditional indoor systems based on dedicated infrastructures.
The research was conducted internally by N&C’s R&D team and has led to the development of experimental models and working prototypes, validated in real-world contexts such as offices and warehouses.
Wi-Fi CSI SENSING
RT INDOOR LOCALIZATION
PRIVACY-BY-DESIGN
Abstract of the scientific publication
‘This paper presents WINP, a low-cost indoor localisation system that leverages existing WiFi infrastructure for real-time asset tracking. By combining Channel State Information (CSI) with Support Vector Machine (SVM) classification and a Kalman Filter, WINP achieves high spatial accuracy and temporal stability. The system uses standard hardware and software-defined radios to extract CSI features, which are processed to estimate position and refined through probabilistic filtering. Extensive experiments in warehouse and office environments demonstrate sub-meter accuracy and robust tracking performance. WINP’s modular architecture enables scalable deployment in dynamic indoor settings, making it suitable for industrial, commercial, and smart building applications.’
Scientific contribution and impact
The WINP project represents a major contribution to the development of WiFi sensing, an emerging technology of considerable international interest, which is being standardised by theIEEE.
For N&C, this research opens up new prospects for innovation, laying the foundations for future products and solutions based on the intelligent use of existing WiFi infrastructures, with applications in areas such as indoor location, smart environments and privacy-preserving presence detection.