GPS-denied Alternative Integrated Navigation (GAIN) System

Recent advancements in non-RF quantum and photonic sensors, database availability and quality, as well as non-linear estimation techniques, have created an opportunity to enable PNT capabilities (including initialization and subsequent continuous operation) without reliance on RF signals in general and GPS/GNSS in particular. To exploit this opportunity, QuNav proposes to develop a GPS-denied Alternative Integrated Navigation (GAIN) system. GAIN integrates three key non-RF enabling technologies with mutually complementary benefits into a joint system architecture:    Inertial Navigation System (INS): INS supports a completely self-contained navigation without relying on external sources. Moreover, recent improvements in inertial sensor quality (particularly, for Micro-Electro-Mechanical Systems or MEMS) enable INS-only functionality over extended periods of time. GAIN uses INS as a core sensor and applies the other two enabling technologies in order to (i) enable the system initialization, and (ii) provide aiding observations to estimate INS error states and adjust navigation outputs thus mitigating the output drift. Celestial-Aided INS: GAIN extends the stellar navigation with angular observations to Low Earth Orbit (LEO) satellites (used for communication, navigation and remote sensing) for improved positioning accuracy.  Map-Matching of Earth Fields’ Anomalies: Navigation on Earth’s fields matches locally observed gradients to a geo-referenced global anomaly map in order to enable localization. A major challenge for this approach is the availability (or lack thereof) of precise maps of the Earth’s fields. To address this challenge, while exploiting richer features of anomaly (more high spatial frequency components) for mounted users at lower altitudes, we propose a two-step strategy. In the first step, we resort to mapping of and positioning relative to field anomaly sources using a simultaneous localization and mapping (SLAM) technique where no anomaly map or only a coarse map exists. Eventually a global map of desired quality is accumulated by piecing together learned local maps, enabling the second step of map-based global navigation. Yet, sparse collection of individual anomaly measurements, though useful in long run to build a dense map, cannot help immediate navigation initially without repeated circling in the same area for sufficient gridding and loop closures. In contrast, we propose to employ analytic inversion of Earth’s field anomalies into discrete sources for SLAM. The sensor fusion architecture utilizes a partitioned estimation approach with separated processing of (i) linear measurements via an extended Kalman filter (EKF), and (ii) non-linear measurements via a Bayesian particle filter.  Phase I will fully develop the navigation mechanization of GAIN and implement it as a post-processing software prototype and initially develop a prototype design approach for Phase II.