Previous Abstract Return to Session B1 Next Abstract

Session B1: Collaborative Navigation Techniques

The PNT Chain: a Collaborative Navigation Architecture for Projecting Accurate PNT Information Across Vast GPS-denied Areas
Samuel Shapero, Daniel Levy, Matthew Lashley, and Mark Smith, Georgia Tech Research Institute
Alternate Number 2

The PNT Chain is a collaborative navigation architecture designed to project accurate position, navigation, and timing (PNT) information across vast distances without GPS. The Chain is composed of a formation of unmanned aerial systems (UASs) equipped with a standard suite of navigation sensors (i.e. an IMU, altimeter, compass, and GPS receiver) and ranging radios. The ranging radios allow the UASs to communicate with their immediate neighbors in the Chain and make range measurements to each other. Range measurements are blended with data from the other sensors onboard the UASs, allowing statistical correlation of the navigation state errors of every UAS in the PNT Chain. This advantageous correlation allows correction of the navigation state errors of all the UASs in the Chain anytime a single member of the Chain receives PNT information. These corrections are propagated via the ranging radios.
Novel distributed estimation algorithms allow the PNT Chain to work in real-time while only requiring communication among neighboring UASs. While all the data from the sensors and ranging radios onboard the UASs would ideally be processed at a single location using a large, centralized Kalman filter, this is infeasible for actual implementation though because of data throughput limitations, single point of failure considerations, and other practical problems. The distributed estimation algorithms developed here for the PNT Chain are based on concepts used in integrated air defense systems (IADS) for fusing target tracks from global networks of radar sites. The distributed algorithms allow the PNT Chain to function in real-time with limited communication resources and only incur a minor loss in performance relative to the optimal, centralized Kalman filter.
A simulation of transoceanic flight demonstrates the ability of the PNT Chain to project accurate PNT information over long distances. Analyses also show that the PNT Chain is able to leverage alternative sources of PNT information (e.g. vision-aiding) and is complementary to other Alt-PNT technologies. The distributed algorithms deliver results almost identical to those of the optimal filter and only require a fraction of the communication bandwidth.



Previous Abstract Return to Session B1 Next Abstract