Dual Frequency Dual Constellation Resilient GNSS Data Collection Prototype

The presence of Radio Frequency Interferences (RFI) / jamming, spoofing and meaconing signals in Global Navigation Satellite System (GNSS) protected bands has increased recently, especially in areas close to conflicts. The presence of those disturbances directly impacts the satellite-based applications.
In the frame of a European Space Agency (ESA) Horizon Europe project aiming to improve GNSS data collection robustness in adverse conditions, a consortium composed by Indra, Syntony and Anywaves and led by Airbus Defence and Space developed and characterized a GNSS Dual frequency Dual constellation Data collection prototype. The main goal is to enhance the integrity, availability and continuity of satellite-navigation based safety-critical applications.
The prototype key components are a 4 elements Controlled Radiation Pattern Antenna (CRPA) developed by Anywaves, two Software Defined Radio (SDR) Receivers L1 and L5 units developed by Syntony and a Core Computer developed by INDRA. The CRPA is composed of 4 GNSS All-Bands antenna. It proposes a small phase centre variation and group delay variation over all the GNSS bands, ensuring a scalability for more GNSS constellations.
The prototype SDR Receiver allows the detection, characterization (including the direction of arrival) and mitigation, for both jammers and spoofers. It incorporates several other functionalities, such as:
- A beamforming technique, to take profit of the number of elements of the CRPA to increase the quality of the received signal
- An auto-calibration, to compensate the gain and phase differences between the RF channels of the SDR
- A CRPA attitude estimation technique.
The prototype covers both L1/E1 and L5/E5a bands of GPS and Galileo constellations. It also includes a set of long RF cables with an indoor rack with several equipment such as a Frequency reference.
After testing the CRPA antenna and the SDR Receiver at unitary level, the full prototype has been integrated and verified by Indra in an incremental manner. This verification included firstly in factory characterisation of the prototype attainable performances in an anechoic chamber and secondly demonstration of the critical functions in a representative environment with real signals in the jammer test campaign 2025 in Norway.
During the anechoic chamber testing performed at Indra’s premises, the prototype was tested with simulated GNSS signals together with jammer signals in order to assess key performances such as jammer detection, jammer angle of arrival and null-forming.
During the jammer live testing, the prototype was tested with real GNSS signals as well as jamming, meaconing and spoofing signals. The impact of these threats on GNSS observables was analysed keeping in mind a perspective of satellite-navigation based safety-critical application.

The widespread use of consumer Unmanned Aerial Vehicles (UAVs) introduces significant conveniences but also critical security threats, particularly regarding unauthorized operations in sensitive areas. Consequently, the development of UAV identification and localization systems has emerged as a vital research area. This paper proposes an integrated system based on passive radar technology, utilizing Universal Software Radio Peripheral (USRP) platforms to passively monitor Radio Frequency (RF) signals in the 2.4 GHz frequency band. By distinguishing the unique RF signatures of various consumer UAVs, the system employs time-frequency spectrogram-based deep learning techniques for rapid identification. Upon successful classification, the target UAV signal is then extracted and processed using the Multiple Signal Classification (MUSIC) algorithm for Direction of Arrival (DoA) estimation. Finally, precise spatial positioning is achieved through a three-station architecture using the Least Squares (LS) algorithm. The core advantage of this technology lies in its passive nature, eliminating the need for additional signal transmitters, which significantly reduces deployment costs and enhances operational concealment.
To classify the UAVs, the system first leverages their unique RF signatures by transforming intercepted signals into time-frequency spectrograms. While most related works rely on general Machine Learning or Deep Learning methods for classification, this paper adopts the You Only Look Once (YOLO) framework. This choice is driven by YOLO's superior real-time object detection capabilities, as well as its well-established architecture and ease of implementation. Crucially, upon target detection, YOLO generates precise bounding boxes that pinpoint the signal's emission time and frequency range. A key advantage of this approach is its ability to mitigate interference. By leveraging YOLO's detection results to selectively extract signal segments from the spectrogram, the system preemptively excludes spectral regions containing severe noise. Consequently, this ensures that the MUSIC algorithm processes only the 'clean' signal components for DoA estimation, thereby preventing the calculation of corrupted signals and significantly enhancing localization precision.
To localize the UAVs, the proposed design adopts a cooperative three-station architecture, where each station is equipped with a six-element Uniform Circular Array (UCA). Following the classification phase, the extracted target signal is processed using the MUSIC algorithm. Unlike conventional beamforming, MUSIC is implemented as a subspace-based super-resolution technique. By exploiting the orthogonality between the signal and noise subspaces, it extracts precise spatial features—specifically, the Angle of Arrival (AoA) and Angle of Elevation (AoE) providing superior accuracy over standard methods. Subsequently, directional information from the three distributed stations is aggregated. To address real-world constraints where DoA lines rarely intersect perfectly due to measurement noise, the LS algorithm is utilized. This method resolves spatial geometry by minimizing triangulation errors, thereby yielding an optimal estimation of the target's three-dimensional coordinates.
The proposed methods make two main contributions: first, the utilization of YOLO to pre-isolate clean signal components effectively mitigates environmental interference, thereby significantly reducing the susceptibility of the MUSIC algorithm to noise during DoA estimation; and second, the leverage of precise time and frequency signatures extracted by YOLO enables the distinction of target signals, allowing the integrated MUSIC and LS algorithms to localize multiple UAVs simultaneously.

Navigation services, nowadays, deploy the state-of-the-art Global Navigation Satellite System (GNSS) for outdoor positioning. The affordability of GNSS receivers, combined with their meter-level positioning accuracy, has made GNSS technology a cornerstone of numerous industries, including automotive, aviation, and transportation. However, because of the long transmission path, the Received Signal Strength (RSS) of GNSS signals at ground level is extremely weak. As a result, GNSS performance is severely degraded in urban areas and canyon-like environments. Moreover, GNSS signals are highly vulnerable to both unintentional interference and deliberate attacks, such as spoofing and jamming.
In this paper, we address the aforementioned shortcomings by exploiting a Dual Polarization Antenna (DPA) processing. Several DPA techniques will be exploited and validated in various testing scenarios. The experimental test campaign will be executed using dual RF-ports recording system with a Radio Frequency Constellation Simulator (RFCS). The RFCS offers 8 RF-ports that are phase-coherent and synchronized in both time and frequency. This enables the RFCS to support spatial and polarization diversity testing.
The experimental scenario exploits GPS and Galileo constellations and various jamming and spoofing conditions. Figure 1 shows the test setup and RFCS capability to simulate both the polarization of signals transmitted from the satellite and that of the receiving antennae. In this specific test, the RFCS was configured to transmit GPS L1 signals with Right-Hand Circular Polarization (RHCP) and Galileo E1 signals with Left-Hand Circular Polarization (LHCP). On the receiving side, two RF ports were used: the first emulating an RHCP antenna (RF1 connected to the Main receiver port) and the second emulating an LHCP antenna (RF2 connected to the Auxiliary receiver port). A Mosaic-go receiver, equipped with two input ports (Main and Aux), was employed to validate the RFCS polarization implementation. The results clearly show that GPS signals were correctly received by the Main port (Fig. 1.b, green box) and Galileo signals by the Aux port (Fig. 1.b, green box).
The primary objective of this paper is to analyse, quantify, and validate the effectiveness of DPA systems in mitigating jamming and spoofing attacks under diverse polarization conditions.

This paper investigates the impact of partial sky blockages on the time and frequency performance of GNSS timing receivers.

Traditionally the recommendation has been to place GNSS timing antennas at locations with as little obstruction of the sky as possible. The reason for this is to maximize the number of visible satellites and to minimize the errors/uncertainties caused by satellite geometry.

Due to the short wavelengths of GNSS, the signal is easily blocked by e.g. terrain and various forms of structures such as buildings. This property of the GNSS signals usually makes optimal antenna placement difficult and limits the number of visible satellites, but this can with advantage be exploited to reduce the strength of any non-local interference and therefore, blocking barriers can act as effective mitigation technique against all forms land based of GNSS jamming.

The static timing application is often constrained in where to place the GNSS antenna, while keeping antenna cable-routing short and practical in a campus environment. Other reasons for the non-optimal antenna placement could be terrain limitations or that the receiver operates in an area with a significant amount of interference coming from a specific direction and therefore the antenna is placed behind a building that protects the antenna from the interfering signal.

Even though GNSS based timing is heavily used in many applications, the impact on restricted sky view is not properly studied. There is only limited knowledge on the impact on metrological time and frequency transfer using undifferenced and differenced observations. The paper revisits Time Dilution of Precision (TDOP) as a metric and assesses the time transfer performance using selective observations in interference free environments.

We reprocessed historic data where satellite observations were removed to simulate varying blocked sky conditions. This reanalysis was validated with measurements done using antennas with physically blocked receptions. The measured values were compared to the output from the operational GNSS receiver(s) as well as the reference time UTC(SP).

The paper concludes with general guidance to users of GNSS timing that operate in locations where the GNSS antenna cannot be placed at the ideal location.

Position, Navigation, and Timing (PNT) services are fundamental to maritime safety and efficiency, with GNSS serving as the primary source of positioning data for both SOLAS and non-SOLAS vessels. However, GNSS vulnerability to interference—whether unintentional or deliberate—poses significant operational risks, including loss of positioning or provision of misleading data. The increasing accessibility of low-cost jamming and spoofing technologies, coupled with advancements in AI-driven signal generation, underscores the urgent need for resilient PNT solutions.
This work introduces a novel approach to resilient PNT for civil maritime applications in the Black Sea and Danube region, addressing GNSS vulnerabilities through an integrated Monitor and Protect approach. Unlike traditional GNSS-reliant systems, this solution leverages a multi-layered strategy combining GNSS radio frequency interference monitoring, alternative positioning through VDES-R / VDES R-Mode and the secure distribution of trusted GNSS Navigation Message data via Application-Specific Messages (ASM) using the AIS/VDES infrastructure and message authentication.
This paper presents an overview of the RIPTIDE Demonstrator trials results with focus on the evaluation of VDES-R (Very High Frequency Data Exchange System Ranging) performance under challenging operational conditions, including both nominal and non-nominal scenarios. The latter scenario analyses the system performance in the presence of non-nominal signals (narrowband, wideband, sweeping) overlapping the VDES-R bands. These conditions were simulated during controlled trials in the Black Sea leveraging a demonstrator system configured with Software Defined Radios and integrated GNSS monitoring capabilities.
The results highlight the systems ability to maintain functional positioning outputs under most interference conditions, with observed degradation in particular cases revealing critical design considerations for operational deployment.
The findings provide actionable insights into the resilience of VDES-R as a complementary PNT source, emphasizing its role in complementing GNSS and supporting maritime navigation continuity in harsh RF environments. Future work will pursue to improve the system response to non-nominal conditions and enhance its robustness.