|dc.description.abstract||A decentralized collaborative navigation algorithm known as inverse covariance intersection (ICI) is studied in the context of a group of vehicles navigating using opportunistic Doppler measurements. Signals of opportunity (SOOPs) have been extensively studied for applications requiring reliable position, velocity, and timing (PVT) information in conditions with potentially degraded GNSS performance. Doppler measurements derived from SOOPs can be used for positioning when GNSS signals are unavailable, but the resulting position estimate accuracy from Doppler-only techniques is unacceptably poor for many use cases. Collaborative techniques can leverage high-quality peer-to-peer range measurements to constrain PVT estimate error growth for each vehicle in a collaborating group.
A navigator employing a tightly-coupled Doppler-inertial extended Kalman filter (EKF) is developed. Its performance is analyzed using Monte Carlo techniques and simulated Doppler measurements from a collection of satellites in low earth orbit (LEO). Peer-to-peer range measurements are integrated using techniques including the well-known covariance intersection (CI), ICI, and a centralized EKF. The performance gains of each method are presented as compared to non-cooperating vehicles. Additionally, the two decentralized navigators are each compared to the centralized navigator, which represents a reasonable best case. The proposed ICI-based navigator is shown using a Monte Carlo test to achieve 62% of the position error reduction of an ideal centralized navigator in the average case, compared to 28% for the well-studied CI-based technique. The proposed ICI navigator is tested with experimentally-collected ranges from ultrawideband transceivers and is shown remain functional in the presence of faulty range measurements.||en_US