Experiences with the OmniSTAR HP Differential Correction Service on an Autonomous Agricultural Vehicle

Stuart Pocknee, Stuart McCarthy, Glen Rains and Craig Kvien

Abstract:	Researchers have been investigating appropriate guidance and navigation methods for robotic agricultural vehicles for decades. Although systems based on machine vision, buried cables, laser range-finding, and other technologies have all shown promise there has been little indication that any would achieve widespread market acceptance. It is possible that GPS will succeed where these other technologies have failed. Systems utilizing RTK GPS for positioning have demonstrated the type of accuracy, robustness, and reliability necessary for autonomous field operations. Multiple prototypes have been built by both commercial and government research groups. Although highly promising, there are still limitations to the technology. Survey grade RTK systems are expensive, technically challenging to operate, and require maintenance of base stations and radio links. An inexpensive, simple, stand- alone DGPS with survey grade accuracies would obviously be an attractive substitute! With this in mind the authors began experimenting with the OmniStar High Performance (HP) correction service in 2002. While not producing survey grade position information, it was thought that the decimeter level positioning information combined with the relatively low cost and obviation of the need for base stations could provide a viable solution for agricultural vehicle autonomy. This paper describes the University of Georgia's ""RowBot"" autonomous tractor and our experiences implementing and testing a guidance/positioning system based on a NovAtel ProPak-LB receiver coupled with OmniSTAR HP correction signals. The paper is divided into three sections. Section 1 describes the autonomous vehicle: it's goals, construction, and technical systems. Section 2 is a qualitative description of our use of the NovAtel/OmniSTAR HP unit. It describes issues surrounding initial acquisition times of precise location coordinates, hardware reliability, technical support, and environmental interferences. In general the unit has performed well despite the rugged environment and demanding conditions. Section 3 is a quantitative analysis of the performance of the NovAtel/OmniSTAR HP combination in both static and dynamic environments. To achieve static results a second NovAtel/OmniSTAR HP unit was placed at a surveyed location for an 11 day period with data logged at 0.2 second intervals. This unit was also used to test the time required for high accuracy position results at startup. To achieve this the unit was power cycled and the .settle time. (time for position readings to return to average high performance distributions) recorded. To quantify dynamic performance we fitted the autonomous vehicle with a Trimble Ag214 RTK receiver for a direct comparison. The vehicle was then driven on a predefined path in the field and results from both positioning systems were compared.
Published in:	Proceedings of the 60th Annual Meeting of The Institute of Navigation (2004) June 7 - 9, 2004 Dayton Marriott Hotel Dayton, OH
Pages:	346 - 353
Cite this article:	Pocknee, Stuart, McCarthy, Stuart, Rains, Glen, Kvien, Craig, "Experiences with the OmniSTAR HP Differential Correction Service on an Autonomous Agricultural Vehicle," Proceedings of the 60th Annual Meeting of The Institute of Navigation (2004), Dayton, OH, June 2004, pp. 346-353.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In