High-Precision RTK Resilience in Asset Tracking
The challenge
Asset trackers operating near steel structures, under bridges, or inside depots frequently face GNSS outages. Pure GNSS solutions cannot guarantee the continuity required for critical logistics and safety use cases. This complex problem demanded a highly robust solution built through deep expertise in product engineering.
What we built
EurthTech delivered a resilient tracking solution combining RTK augmentation for high-precision with an AI-powered embedded system for navigation fallback. The core technology is sensor fusion, continuously blending data from multi-constellation GNSS, IMUs (Inertial Measurement Units), and opportunistic RF fixes (BLE/LoRa). This Edge AI embedded system ensures seamless tracking continuity and establishes a crucial layer of Predictive maintenance AI IoT.
How it works
The embedded systems development focused on a tightly-coupled IMU and a server-side GeoAI fusion engine utilizing an Extended Kalman Filter (EKF). This engine blends noisy GNSS/RTK fixes with IMU dead-reckoning during outages. Opportunistic RF fixes (BLE fingerprinting, cell-tower trilateration) and map-matching logic, leveraging AI GIS analytics, snap tracks to plausible corridors. The system uses an adaptive restart policy to blend the dead-reckoned state with freshly-resolved RTK fixes, ensuring a graceful recovery.
Results
Field trials demonstrated maintained tracking continuity. The hybrid stack reduced "lost-location" events by 90% and improved end-to-end chain-of-custody confidence. Mean gap duration in asset track continuity dropped from several minutes to sub-minute levels in mixed environments.
Scalability & GIS Context: Deployment requires precise RF planning and managing the RTK reference network, a core competence of our GIS consulting company. The map-matching and coordinated RF corrections are delivered through comprehensive geospatial engineering services. This high-integrity, continuous location data is essential for building a reliable, actionable Digital twin smart city model.