Agentic data science with the Data Agent Kit
cloud.google.com·7h ago
AI Research
arXiv CS.AI·25d ago·5 min read
Physics-Constrained Digital Twins for Sensor Integrity in Urban Pedestrian Flow: Detecting Stealthy False Data Injection with Conformal Guarantees
TL;DR
A physics-constrained digital twin was developed to enhance the integrity of urban pedestrian flow data by detecting stealthy false data injections. This approach demonstrated a significant reduction in the impact of compromised sensors, achieving a 0.54 attack margin against a single device and 0.19 with multiple devices compromised.
✦ Why It Matters
Implement physics-constrained models to enhance sensor data integrity in urban planning today.
Key Takeaways
- Adopt physics-based models for urban sensor networks.
- Monitor sensor integrity to prevent data manipulation.
- Utilize flow conservation principles in data analysis.
- Evaluate the impact of compromised sensors on data accuracy.
How It Works
The digital twin estimates pedestrian flows using a directed street graph and incorporates a learned graph-localized gain. It ensures flow conservation by linking metered and unmetered segments, allowing for accurate detection of anomalies in sensor data.