Abstract
Positioning, Navigation, and Timing (PNT) systems are of critical importance to modern applications, ranging from the positioning of smartphones to the timing of financial and energy infrastructure. However, reliance on Global Navigation Satellite Systems (GNSS) renders PNT vulnerable to signal weaknesses, jamming, and limitations in urban and indoor environments. While terrestrial network-based positioning has advanced from 1G to the emerging 6G, no single solution meets all PNT needs. Multi-technology fusion offers a promising approach but poses challenges due to varying frequencies, error characteristics, and the need for adaptive algorithms. Traditional Kalman filters, widely used in data fusion, struggle with static noise covariance assumptions, leading to suboptimal performance in highly dynamic environments. To address these limitations, this research proposes an AI-driven approach to enhance positioning accuracy, resilience, and adaptability by dynamically tuning noise covariance parameters and improving error compensation. This research aims to advance state-of-the-art positioning algorithms by developing a robust multi-epoch positioning framework that effectively fuses heterogeneous PNT data sources. The proposed AI-driven fusion approach will significantly improve positioning accuracy, availability, and resilience in diverse operational environments, addressing the challenges associated with next-generation PNT systems.
Researcher(s)
Research team(s)
Project type(s)