VAMPIR – Verifiable AI/ML for Positioning with Improved Reliability is an innovative project under ESA's Navigation Innovation and Support Program (NAVISP). NAVISP is designed to drive innovation and competitiveness within the European Positioning, Navigation, and Timing (PNT) landscape, a vital area for modern society where PNT data is crucial for enhancing connectivity and automating processes. The program not only focuses on satellite navigation signals and data but also encourages the development of innovative technologies that can complement, enhance, or even replace current PNT systems, thereby fostering industrial growth and creating competitive capabilities across Europe.
The VAMPIR project aims to push the boundaries of PNT technology by integrating conventional PNT methods, such as Global Navigation Satellite Systems (GNSS) and inertial sensors, with advanced Artificial Intelligence (AI) and Machine Learning (ML) techniques. This combination is designed to significantly improve the reliability and accuracy of PNT systems, reducing the risk of large positioning errors that could compromise critical applications.
A key aspect of VAMPIR is the development of a Verification Engine, which serves as a safeguard for the AI-enhanced PNT solutions by using graph reasoning. This engine employs formal verification techniques to run consistency checks on the system's outputs, ensuring that any errors are promptly detected and corrected. By enhancing the verifiability and integrity of AI-driven PNT systems, VAMPIR aims to offer a more secure and dependable solution for positioning applications, especially in scenarios where precision and reliability are paramount.
To demonstrate the practical potential of these advancements, the project will deliver a working demonstrator within the automotive sector, where precise and reliable navigation is essential for autonomous vehicles and advanced driver assistance systems. The demonstrator will showcase how AI-enhanced PNT solutions can improve vehicle navigation, integrating camera-based imagery with PNT data to achieve a highly accurate, verified positioning system.
IncQuery as the consortium leader, contributes through technology transfer, tool integration expertise, and commercialization strategies. This collaboration strengthens our innovation edge in applying formal methods (e.g., Dynamic Verification Toolkit ) for V&V of complex systems.
Budapest University of Technology and Economics (BME): Two research groups from BME contribute to the VAMPIR project:
This project is supported by ESA’s NAVISP program, which fosters innovative PNT solutions across Europe.
