064 - Block-box for an optimised GNSS spectrum monitoring network using AI
There is an abundance of GNSS receivers deployed around the world providing measurements and PNT information, which are currently combined into networks for a variety of GNSS processing and monitoring applications. However, these receivers remain vulnerable and cannot be used to isolate all sources of error on all GNSS signals and constellations due to the processed nature of the receiver output, which is receiver-specific and sometimes also due to the encrypted nature of the transmitted GNSS signals.
To overcome these difficulties, it is proposed to develop the Block-box, an external RF2RF device targeting the enhancement of any COTS GNSS receiver.
It can be tuned to a variety of applications at user, downstream services or system level supporting monitoring functions and real-time signal cleaning.
It is based on AI helping to detect and categorise signal and system anomalies and interference. The starting TRL is low and expected to be raised during the activity taking past and current TRP activities (TERMINATE and AIMGNSS) as input. These detection capabilities will aim to be remotely upgradeable with newly trained models in order to, for example, keep up with constantly expanding and increasingly sophisticated RFI/jammer profiles.
Observations are based on measurements directly taken in the Block-box from its GNSS spectrum Intermediate Frequency (IF) samples and on measurements coming from the associated COTS GNSS receiver and enhanced by the Block-box. Rejection of interfering signals is performed on the same RF signal injected in the COTS GNSS receiver.
In order to support monitoring functions, Block-box will buffer recent IF samples to capture the period around an event detection. These samples can be stored and used later, for example, to test and tune new receiver algorithms.
Extra parameters for integrity are also targeted in the block-box by enhancing the raw measurements of the GNSS receiver and the detection of satellite signal anomalies, e.g evil waveforms, exploiting the potential to network several devices together.
The objectives of the proposed activity are to investigate, prototype and validate an RF2RF GNSS receiver enhancement device, the Block-box. This device uses a GNSS spectrum sampler in order to monitor and clean the incoming signal, making use of AI/ML techniques.
The tasks to be performed include:
- Reviewing state-of-the-art monitoring and signal cleaning techniques using AI;
- Defining of relevant use-cases, environments and concepts of operation;
- Designing, development and breadboarding of a not full-scale breadboard model of the block-box device and SW algorithms;
- Verifying and demonstrating the basic functional performance. The verification includes tests in laboratory environment and performance verification through testing in the relevant environment, subject to scaling effects.
The main output of the activity will consist of:
- Block-box breadboard verified in relevant environment (TRL 4-5)
- Documentation including:
- Preliminary definition of performance requirements and of the relevant environment (Preliminary technical requirements specification)
- Analysis report for technology associated with critical functions
- Preliminary design of the element, supported by appropriate models for the verification of the critical functions (Preliminary design definition file)
- Preliminary design justification file including: identification of computational design methods and tools; analysis of scaling effects; breadboard definition for the verification of the critical function of an element;
- Test plan and reports