ANTIFERENCE

New concept for evolutive mitigation of RFI (Radio Frequency Interference) to GNSS

Aims

The main envisaged objectives of this project are:

• Investigation of the feasibility of flexible and reconfigurable Digital Signal Processing (DSP) techniques for GNSS interference mitigation.
• Investigation of the methods for the identification of new GNSS interference and fingerprint extraction, allowing to reconfigure the DSP to effectively mitigate them.
• Validate the proof of concept (PoC) via breadboarding and demonstration.

Brief Description

The project targets the detection and mitigation of RF interference using methods from Artificial Intelligence, whereby the detection and mitigation techniques will be derived using Machine Learning approaches.

Key characteristics of the project:

• Trained, tested, and trade-off analysis of different models developed for the selected use-cases
• Investigation, simulation and testing of smart jamming
• Generation of RF fingerprints to build-up a database
• Collection and pre-processing of real world data (When existing datasets do not cover scenarios of interest, such as spoofing)
• Applying mitigation strategies (e.g. adaptive filters) based on ML (Machine Learning) filter parameter prediction
• Benchmarking the developed algorithms w.r.t. COTS GNSS receiver

Based on the technique design and on the use-cases identified, two different high-level system concepts will be investigated – focusing on both, a centralized (cloud-based) approach and a decentralized approach. Within the centralized/cloud-based approach the RF samples are uploaded to a central facility which performs the feature extraction and fingerprint generation. The receiver (user equipment) will only contain the adaptive mitigation component. Within the decentralized approach each receiver performs independently the feature extraction and fingerprint generation but is able to share the fingerprints with other receivers.