




NADODE develops alternative navigation methods for drones operating in GNSS-denied areas, essential for defense and emergency response.
NADODE Autonomous Navigation Beyond GNSS
The escalating threats posed to Global Navigation Satellite Systems (GNSS) by deliberate interference, encompassing interference, jamming, and spoofing, necessitate a dual-pronged approach. Firstly, it requires the consideration of employing multiple sensors or sensor fusion technologies as viable alternatives to GNSS, especially in contexts where safety and military operations are at stake. Secondly, it demands the development of advanced methods with heightened sensitivity to detect and pinpoint the sources of interference, including Unmanned Aerial Vehicles (UAVs), in real-time to assess instances of compromised or incapacitated GNSS navigation. Notably, this includes scenarios where drones are compelled to rely on alternative navigation methods due to GNSS impairment, with the aim of identifying and neutralizing interference from both ground-based and aerial sources.
UAV against Interferer
The primary objective of this project is twofold. Firstly, it seeks to identify the most suitable non-GNSS navigation sensors or sensor fusion approaches for both civilian and military applications. This is accomplished through extensive simulations, the development of criteria matrices tailored to specific use cases, SWOT analyses, and a thorough review of existing field reports in the literature.
The second key objective centers on the implementation of more sensitive interference detection algorithms, with a particular focus on the Karhunen-Loève transformation (KLT) in software. This method is rigorously tested via simulations involving various GNSS interference scenarios and is compared against conventional approaches such as the Fast Fourier Transform (FFT). Ultimately, it is intended to undergo practical testing.
Karhunen-Loève transformation proof of concept
The third objective entails the realization of a Proof-of-Concept by deploying a drone equipped with a demonstrator for genuine non-GNSS navigation and KLT-based interference detection. This milestone is set to occur for the first time at the Seetaler Alpe military training area (TÜPL), utilizing the LOKI signal generator developed by IGASPIN as the jamming and spoofing source.
Additionally, the project includes the application of a specific non-GNSS navigation method. In this context, image-based localization utilizing passive sensors is chosen and implemented. The fundamental principle behind visual navigation involves comparing captured image data with an existing visual representation of the environment, or one generated concurrently. This alternative navigation method is evaluated using recorded data, with post-processing conducted on actual recorded data.
In sum, this project contributes significantly to enhancing the security and availability of critical infrastructure and services that rely on GNSS technology.
The project was partly funded within the FORTE program by the österreichische Forschungsförderungsgesellschaft mbH (FFG), Vienna and financed by the Austrian Bundesministerium für Finanzen (BMF).
In collaboration with
- JOANNEUM RESEARCH Forschungsgesellschaft mbH
- Bundesministerium für Landesverteidigung (BMLV)
Do you need specific information?
Get all the answers you need. Send us your message now!
Related products
INDALOS
GNSS Interference Detection & Localization
Protect your GNSS systems from interference! INDALOS is an AI-powered monitoring solution that detects and localizes GNSS jamming and spoofing in real time. With networked sensors and advanced analytics, it provides precise threat classification and immediate counteraction to prevent signal manipulation.
LOKI
GNSS Interference Simulator
LOKI is your advanced test environment for GNSS interference resilience. Simulate real-world jamming and spoofing attacks under controlled conditions, assess system vulnerabilities, and optimize performance against interference threats. Trusted by research institutions, defense organizations, and the automotive industry.