Software-Defined VTL GNSS Receiver with AI-enhanced Resilience in High Dynamics and Stringent Conditions

Global Navigation Satellite Systems (GNSS) have become the primary means of navigation and source of Position,
Navigation and Timing (PNT) information for almost all modes of transport, for general navigation and for timing
purposes. Yet, users operating in urban canyons, or within dense vegetation may suffer from a limited view of the
satellites and multipath conditions, thus encountering difficulties in providing reliable positioning solutions. In such
scenarios, the benefits of Vector Tracking Loops (VTL) in comparison to conventional Scalar Tracking Loops (STLs)
can be exploited to tackle high user dynamics and stringent positioning requirements, thereby countering reflected,
weak or distorted signals. The advantages of VTL help PNT performance by utilising navigation filter outcomes to
compensate for track phase, frequency, and delay offsets. Nevertheless, the VTL is vulnerable to signal degradation
due to its dependency on individual tracking channels. Hence, fusion with an Inertial Measurement Unit (IMU) and
Receiver Autonomous Integrity Monitoring (RAIM) techniques becomes advantageous under the stringent conditions
affected by signal reflection, multipath, and attenuation. To mitigate the challenges of operating under challenging
environments and integrating multiple AI implementations into software-defined GNSS receivers, this paper presents a
new software-defined GNSS architecture incorporated with modern techniques. Specifically, the VTL development
facilitates resilience against Non-Line of Sight (NLOS) effects. The adoption of the Ultra-tightly Coupled (UTC) fusion
mechanism enhances high-dynamic compatibility. The AI feature strengthens detector accuracy and facilitates enhanced
Extended Receiver Autonomous Integrity Monitoring (eERAIM) for fault detection and exclusion. The architecture
performance is evaluated through functionality tests at component level by using a high-fidelity GNSS simulator to
generate synthetic datasets. Using indicators in terms of availability and integrity to assess performance, the proposed
VTL design offers significant advantages, outperforming STL at a cost of increased computational complexity and IMU
integration. Using the integration VTL with AI, our proposition presents the potential to improve resilience in GNSS
receivers.