Information Density Metric for AI-Driven Virtual Sensing
A recent study presents Information Density as a measurable parameter aimed at enhancing sensor placement and facilitating AI-based virtual sensing. This framework leverages spatial, temporal, and inter-modal relationships between sensor signals to execute sensing functions without the need for physical sensors. Two specific metrics—Phase in Eigen Space and Mutual Information—are employed to assess information density for the best sensor arrangement. Validation of this method was conducted using actual data from the smart city infrastructure in Madrid.
Key facts
- arXiv:2605.08180v1
- Published on arXiv
- Introduces Information Density as a quantitative metric
- Framework leverages spatial, temporal, and inter-modal correlations
- Two measures: Phase in Eigen Space and Mutual Information
- Validated using real-world data from Madrid's smart city infrastructure
- Aims to address challenges in storage, transmission, and real-time processing of IoT data
- Contrasts with traditional compressive sensing and machine learning-based compression
Entities
Institutions
- arXiv
Locations
- Madrid