Abstract
Abstract. Unmanned Aerial Vehicle (UAV) technology is now commonplace in many defence and civilian environments. However, the high cost of owning and operating a sophisticated UAV has slowed their adoption in many commercial markets. Universities and research groups are actively experimenting with UAVs to further develop the technology, particularly for automated flying operations. The two main UAV platforms used are fixed-wing and helicopter. Helicopter-based UAVs offer many attractive features over fixed-wing UAVs, including vertical take-off, the ability to loiter, and highly dynamic flight. However the control and navigation of helicopters are significantly more demanding than those of fixed-wing UAVs and as such require a high bandwidth real-time Position, Velocity, Attitude (PVA) navigation system. In practical Real-Time Navigation Systems (RTNS) there are delays in the processing of the GNSS data prior to the fusion of the GNSS data with the INS measurements. This latency must be compensated for otherwise it degrades the solution of the navigation filter. This paper investigates the effect of latency in the arrival time of the GNSS data in a RTNS. Several test drives and flights were conducted with a low-cost RTNS, and compared with a high quality GNSS/INS solution. A technique for the real-time, automated and accurate estimation of the GNSS latency in low-cost systems was developed and tested. The latency estimates were then verified through cross-correlation with the time-stamped measurements from the reference system. A delayed measurement Extended Kalman Filter was then used to allow for the real-time fusing of the delayed measurements, and then a final system developed for on-the-fly measurement and compensation of GNSS latency in a RTNS.
Highlights
In real-time implementations of GNSS/INS looselycoupled systems the GNSS measurements suffer from a delay such that the measurements must be fused late with respect to the state of the system
This paper focuses on developing a reliable method for estimating GNSS latency in loosely-coupled GNSS/INS systems
It is noted that research conducted prior to the development of the real-time GPS/INS navigation systems tend to overlook the issue of delayed measurements
Summary
In real-time implementations of GNSS/INS looselycoupled systems the GNSS measurements suffer from a delay such that the measurements must be fused late with respect to the state of the system. This is because the inertial loop in a navigation filter typically runs at a much higher rate than the rate at which GNSS measurements come in and several inertial epochs will have passed before a GNSS measurement is ready for fusion.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
