top of page

Trusted Positioning Receives Best Presentation Award at ION GNSS+ 2026!

11 minutes ago
3 min read

When it comes to positioning technology, accuracy matters. And at ION GNSS+ 2026, our research hit the mark.


Trusted Positioning was honoured with the Best Presentation Award for Session B5b, recognizing our team's work on improving the accuracy and reliability of autonomous vehicle positioning.


ION GNSS+ is an annual conference hosted by The Institute of Navigation (ION) that brings together international leaders in positioning, navigation, and timing (PNT) to share their latest research, introduce new technologies, and share ideas. 


This year, Trusted Positioning submitted two papers to ION GNSS+. The first discussed Vision-Based 3D Map Matching with Tight Integration of Vision/Radar/INS/GNSS for Positioning Autonomous Systems, and the second discussed AI-Enhanced GNSS Quality Classification to Improve the Accuracy and Reliability of Autonomous Vehicles and won the award for best presentation for session B5b!


Our team is so proud of this achievement and of all the hard work that has led to this moment. At Trusted Positioning, we are constantly innovating our technology and making advancements in sensor fusion and positioning software.


Follow us on LinkedIn and subscribe to our newsletter so you never miss out on our latest product launches, upcoming trade shows, and more!


Implications of this Research 


Standard GNSS positioning is effective in environments with unobstructed sky; however, even RTK GNSS performance deteriorates significantly in complex settings such as urban canyons and tunnels due to multipath interference and signal blockages. To address this challenge, this paper introduces a novel AI framework that classifies GNSS positioning quality to dynamically adapt the integrated navigation solution. Experimental results demonstrate that this approach significantly outperforms conventional quality assessment methods, resulting in meaningful reductions in positioning error and enhanced system integrity during severe GNSS degradation. Being that Trusted Positioning’s embedded software is sensor-agnostic; these improvements to the accuracy of a positioning solution are achieved without additions or modifications to hardware, which facilitates easy integration and deployment at scale. 


Read the abstract of this award-winning presentation below.


AI-Enhanced GNSS Quality Classification to Improve the Accuracy and Reliability of Autonomous Vehicles


Abstract


Ensuring continuous, reliable, and highly accurate positioning is a core requirement for autonomous vehicle navigation, where lane-level decision-making and vehicle control depend on decimeter-to-centimeter level precision. To maintain this performance across diverse environments, navigation systems typically fuse GNSS and INS. However, RTK GNSS performance deteriorates significantly in complex settings such as urban canyons and tunnels due to NLOS signals, multipath interference, and signal blockages. A critical vulnerability is that lower-end, automotive-grade GNSS receivers routinely underestimate their position standard deviation measurements in these scenarios. When these overconfident, faulty measurements are integrated into the INS/GNSS navigation filter, they heavily bias and degrade the overall vehicle state estimation.


 To address this challenge, this paper introduces a novel cross-vendor AI framework that classifies GNSS positioning quality to dynamically adapt the integrated navigation solution. The proposed method partitions continuous GNSS data streams into overlapping segments using motion-aware windowing strategies to preserve temporal continuity. These segments are processed by the advanced deep learning technique proposed in this paper, to capture complex nonlinear temporal dynamics and assign a reliable quality level and confidence factor to each segment. Crucially, the framework achieves hardware-agnostic scalability by relying exclusively on standardized GNSS observables and derived motion features rather than proprietary, manufacturer-specific diagnostics. The introduced method was trained and validated on large-scale multi-sensor datasets spanning multiple cities and geographical settings, using a wide range of GNSS modules from consumer-grade chipsets to high-precision receivers. Experimental results demonstrate that this approach significantly outperforms conventional quality assessment methods, yielding meaningful reductions in positioning error and enhanced system integrity during severe GNSS degradation events while lowering the hardware barrier to entry for wide-scale autonomous deployment.


We invite you to read the full paper on AI-Enhanced GNSS on the ION website by clicking here: https://www.ion.org/gnss/abstracts.cfm?paperID=16947



Glossary


GNSS: Global Navigation Satellite System

INS: Inertial Navigation System 

NLOS: Non-Line of Sight 

RTK: Real-Time Kinematics



bottom of page