GPS-IMU

Development of a High Accuracy Pointing System for Maneuvering Platforms

When it comes to providing attitude or designing a pointing system — say, for controlling an airborne antenna or guiding a ship on maneuvers, GPS can use some help. That’s where inertial measurement technology enters the picture. Navigation-grade IMUs can handle the job, but researchers at SRI were challenged to see if a less expensive MEMS design could provide suitable accuracy. Turns out it could.

Joseph M. Strus, Michael Kirkpatrick, and James W. Sinko

» Download this article (PDF)

SRI International (SRI) has recently addressed the requirements of pointing systems for a variety of maneuvering platforms. These platforms include airborne systems (unmanned aerial vehicles, aircraft), land vehicles (tanks, HUMVEES), and marine vessels.

Our primary goal was to obtain 0.1-degree pointing accuracy. To achieve this, we considered several design options. A stand-alone navigation grade inertial measurement unit (IMU) seemed too expensive and heavy but has a clear advantage by being more immune to GPS outages. A magnetic compass–based solution appeared too problematic due to calibration and accuracy issues.

After other design trades were reviewed, we limited the path forward to tactical grade IMUs combined with GPS. Several different IMUs were then evaluated for integration into a flexible software package previously developed at SRI for position and attitude tracking of large parachute pallet loads.

A secondary goal was to establish a truth system to verify pointing accuracy of the developed system. The criteria that we set for the truth system were approximately 0.06 degree for kinematic applications and 0.02 degree for static applications. Moreover, we wanted all biases between the units under test and the truth system to be less than 0.01 degree.

Providing truth at this level of accuracy presents difficulties, however. Optical systems can easily attain this level of accuracy for static tests but are difficult for dynamic tests.

A stand-alone GPS attitude system works well for kinematic tests, but the static accuracy requirement would need too long of a baseline to be portable. Ultimately, a hybrid system was developed using both optical and GPS methods.

The first part of this article presents the component analysis and differences for the MEMS IMU versus the tactical grade unit. Then we discuss the design and architecture for the system and the associated GPS/INS navigation processing software. Next we discuss implementation differences for the various components.

Following those sections, we consider the truth systems developed at SRI. Finally, we discuss the tests performed, truth data analysis methodology, and results.

This SRI initiative has led to the implementation of GPS/IMU systems on a variety of platforms. . .

Conclusions
With suitable dynamics, both varieties (fiber-optic and MEMS) of IMU/GPS combinations were capable of providing an azimuth to within at least 0.06° 1 σ. Furthermore, the Allan variance analysis accurately predicted the azimuth drift performance of the IMU systems.

Additional testing on the FOG units showed azimuth to be determined faster and more accurately with RTK data than with L1 data. The telescopic sight proved to be a convenient way of testing for static cases. The long-boom GPS attitude system, coupled with averaging, appears to give very good testing accuracy during dynamics.

Acknowledgment: We wish to thank Patrick Weldon of Honeywell for lending us on short notice the MEMS unit used in our tests.

(For the complete article, including figures, graphs, and additional resources, download the PDF version at the link above.)

Manufacturers

The fiber-optic tactical grade IMU described in this article was the A=2 model of the Litton LN-200 (311875-240207) manufactured by Northrop Grumman, Woodland Hills, California, USA. The MEMS IMU was the HG1900 BA99 by Honeywell Aerospace Electronic Systems, Minneapolis, Minnesota, USA. The INS systems incorporated an OEMV receiver from NovAtel, Inc., Calgary, Alberta, Canada, using the OmniSTAR satellite-based augmentation service, Houston, Texas, running in the VBS/XP modes.

Author Profiles

Joseph M. Strus <joseph.strus@sri.com> is a systems analyst at SRI International, Menlo Park, California, where he has worked on precision navigation applications since 2001. His interests are GPS, GPS/INS, and estimation theory.

Previously, he was a GPS systems engineer with the Government Systems Division of Rockwell Collins. Strus received his B.S. and Ph.D. (mathematics) from the University of Illinois at Urbana-Champaign and holds several GPS-related patents.

Michael Kirkpatrick <michael.kirkpatrick@sri.com> is a senior research engineer at SRI International, where he has been involved in hardware and software systems design and development since 1985.

James W. Sinko is a principal engineer at SRI International. He received his B.S. (engineering science) and M.S.E.E. from Stanford University, and his Ph.D. (E.E.) from the University of Rochester. Sinko has been with SRI since 1967, working with radar and aircraft systems. For the last 13 years, he has been working with precision GPS for military and civil applications.